CN104077768B - Method and device for calibrating fish-eye lens radial distortion - Google Patents

Abstract

The embodiment of the present invention discloses a calibration method and calibration device for the radial distortion of a fisheye lens, wherein the method includes the following steps: acquiring a calibration image containing at least two sets of spatially parallel straight lines; The spatially parallel straight lines are fitted to the center-point collinear arc group mapped on the distorted image; the distortion calibration model is generated according to the fitting processing results, which can improve the accuracy of the calibration results of the distortion calibration model.

Description

Calibration method and calibration device for radial distortion of fisheye lens

technical field

The invention relates to the field of electronic technology, in particular to a calibration method and a calibration device for radial distortion of a fisheye lens.

Background technique

With the continuous development and improvement of electronic technology, cameras, mobile phones and other terminal devices with camera and video functions have been widely used in people's daily life. In order to increase the expression of captured images, users usually use some special effects Lenses, such as: fisheye lens, tilt-shift lens, macro lens, etc., increase the effect and variety of image shooting.

The fisheye lens is an extreme wide-angle lens. In order to achieve the maximum photographic viewing angle, the diameter of the front lens protrudes toward the front of the lens in a parabolic shape. The angle of view of the fisheye lens strives to reach or exceed what the human eye can see. scope. Therefore, there is a big difference between the fisheye lens and the real world scene in people's eyes. However, there are many types of distortion in the images captured by the fisheye lens, among which the radial distortion is considered to be the main factor affecting the image quality, because the traditional radial distortion calibration process is limited to fitting the arc corresponding to a single line to The distortion parameters are obtained to generate a distortion calibration model, which leads to a low accuracy rate of the calibration results of the distortion calibration model, and cannot well solve the problem of radial distortion of the fisheye lens.

Contents of the invention

Embodiments of the present invention provide a calibration method and a calibration device for radial distortion of a fisheye lens, which can improve the accuracy of calibration results of a distortion calibration model.

In order to solve the above technical problems, the first aspect of the embodiment of the present invention provides a method for calibrating the radial distortion of a fisheye lens, which may include:

Obtain a calibration image comprising at least two sets of spatially parallel straight lines;

performing fitting processing on the group of collinear arcs with the center of the circle mapped on the distorted image by each of the at least two groups of spatially parallel straight lines;

A distortion calibration model is generated according to the fitting processing results.

Based on the first aspect, in the first feasible implementation manner of the first aspect, any set of spatially parallel straight lines in the at least two groups of spatially parallel straight lines are mapped to a group of collinear arcs with centers of circles on the distorted image, so Any straight line in any set of spatially parallel straight lines has a corresponding arc in the center-collinear arc group mapped on the distorted image.

Based on the first aspect or the first feasible implementation manner of the first aspect, in the second feasible implementation manner of the first aspect, each set of spatially parallel straight lines in the at least two groups of spatially parallel straight lines is The center collinear arc group mapped on the distorted image is used for fitting processing, including:

Using a preset non-linear optimization algorithm to calculate the common parameters and proportional coefficients of the center collinear arc group mapped on the distorted image by each group of spatial parallel straight lines in the at least two groups of spatial parallel straight lines;

Using the public parameters and the proportional coefficient to calculate and obtain a fitting processing result, the fitting processing result includes a parametric representation of each arc in each of the center collinear arc groups and each of the center collinear circles The two public intersection points corresponding to the arc group;

Wherein, the public parameters include the midpoint coordinates of the line between the two public intersections and the vector from the midpoint to any intersection point of the two public intersections, and the proportional coefficient includes the collinear The preset coefficient corresponding to each arc in the arc group.

Based on the second feasible implementation manner of the first aspect, in the third feasible implementation manner of the first aspect, the preset nonlinear optimization algorithm is used to calculate each of the at least two sets of spatially parallel straight lines The public parameters and scale coefficients of the group space parallel lines mapped on the distorted image and the collinear arc group, including:

Obtaining the square distance minimization equation corresponding to the group of collinear arcs with the centers of each group of the at least two groups of spatially parallel straight lines mapped on the distorted image;

A preset nonlinear optimization algorithm is used to calculate the square distance minimization equation to obtain the common parameters and proportional coefficients of each of the collinear arc groups with centers.

Based on the second feasible implementation manner of the first aspect, in the fourth feasible implementation manner of the first aspect, the generation of the distortion calibration model according to the fitting processing result includes:

Obtain the coordinates of the distortion center point and the distortion parameters according to the fitting processing result;

A distortion calibration model is generated by using the coordinates of the distortion center point and the distortion parameters.

Based on the fourth feasible implementation manner of the first aspect, in the fifth feasible implementation manner of the first aspect, the obtaining the distortion center point coordinates and distortion parameters according to the fitting processing result includes:

Obtain the straight line equations of the two common intersection points corresponding to the collinear arc groups of the respective circle centers, and determine the intersection coordinates of all straight line equations as the distortion center point coordinates;

The distortion parameters are obtained according to the coordinates of the distortion center point and the parametric representation of any circular arc in any circle center collinear circular arc group.

Based on the third feasible implementation manner of the first aspect, in the sixth feasible implementation manner of the first aspect, the square distance minimization equation is:

Among them, m represents the number of points fitting the i-th circular arc, |Q _i,j o _i | represents the distance between point Q _i,j on the i-th circular arc and the center o _i of the circle, r _i represents The radius of the i-th arc, n represents the number of arcs in a center collinear arc group, (p _x , p _y ) represents the midpoint P of the line between the two public intersection points M and N The coordinates of the point, (t _x , _ty ) represents the vector from the P point to the M point or N point, and (k ₁ ... k _n ) represents the proportional coefficient of each arc in the n arcs;

The preset nonlinear optimization algorithm is LM (Leven-Marquardt) algorithm.

Based on the fifth feasible implementation manner of the first aspect, in the seventh feasible implementation manner of the first aspect, it is assumed that the parametric representation of any arc in any circle center collinear arc group is expressed as The coordinates of the distortion center point are (x _d0 , y _d0 ), then the distortion parameters are:

where A, B and C are the parameters in the parametric representation of any arc.

Based on the seventh feasible implementation manner of the first aspect, in the eighth feasible implementation manner of the first aspect, the distortion calibration model is:

Among them, b=(x _d , y _d ) is the coordinate point of the distorted image, a=(x _u , y _u ) is the coordinate point of the image after distortion correction, and r _d is the Euclidean distance between b and the coordinate of the distortion center point.

The second aspect of the embodiment of the present invention provides a calibration device for radial distortion of a fisheye lens, which may include:

A calibration image acquisition module, configured to acquire calibration images comprising at least two sets of spatially parallel straight lines;

An arc fitting module, configured to perform fitting processing on the group of collinear arcs with the center of the circle mapped on the distorted image by each group of spatial parallel straight lines in the at least two groups of spatially parallel straight lines;

The model generation module is used to generate a distortion calibration model according to the fitting processing result.

Based on the second aspect, in the first possible implementation manner of the second aspect, any set of spatially parallel straight lines in the at least two groups of spatially parallel straight lines are mapped to a group of collinear arcs with centers of circles on the distorted image, so Any straight line in any set of spatially parallel straight lines has a corresponding arc in the center-collinear arc group mapped on the distorted image.

Based on the second aspect or the first possible implementation manner of the second aspect, in the second feasible implementation manner of the second aspect, the arc fitting module includes:

A parameter coefficient calculation unit, configured to use a preset nonlinear optimization algorithm to calculate the public parameters and proportional coefficients of the center collinear arc group mapped on the distorted image by each set of spatial parallel straight lines in the at least two groups of spatial parallel straight lines ;

A result acquisition unit, configured to calculate and obtain a fitting processing result by using the public parameters and the proportional coefficient, the fitting processing result includes a parametric representation of each arc in each of the center-center collinear arc groups and each Two public intersection points corresponding to the group of collinear arcs with centers of circles;

Wherein, the public parameters include the midpoint coordinates of the line between the two public intersections and the vector from the midpoint to any intersection point of the two public intersections, and the proportional coefficient includes the collinear The preset coefficient corresponding to each arc in the arc group.

Based on the second feasible implementation manner of the second aspect, in a third feasible implementation manner of the second aspect, the parameter coefficient calculation unit includes:

An equation obtaining subunit, configured to obtain a square distance minimization equation corresponding to a group of collinear arcs with center points mapped on the distorted image for each group of spatially parallel straight lines in the at least two groups of spatially parallel straight lines;

The parameter coefficient acquisition subunit is used to calculate the square distance minimization equation by using a preset nonlinear optimization algorithm, and acquire the common parameters and proportional coefficients of each of the collinear arc groups with centers of circles.

Based on the second feasible implementation manner of the second aspect, in the fourth feasible implementation manner of the second aspect, the model generation module includes:

A coordinate parameter acquiring unit, configured to acquire the coordinates of the distortion center point and the distortion parameters according to the fitting processing result;

A model generation unit, configured to generate a distortion calibration model by using the coordinates of the distortion center point and the distortion parameters.

Based on the fourth feasible implementation manner of the second aspect, in the fifth feasible implementation manner of the second aspect, the coordinate parameter acquisition unit includes:

The coordinate obtaining subunit is used to obtain the straight line equations of the two common intersection points corresponding to the collinear arc groups of the respective circle centers, and determine the coordinates of the intersection points of all the straight line equations as the coordinates of the distortion center point;

The parameter acquisition subunit is configured to acquire distortion parameters according to the coordinates of the distortion center point and the parametric representation of any arc in any circle center collinear arc group.

Based on the third feasible implementation manner of the second aspect, in the sixth feasible implementation manner of the second aspect, the square distance minimization equation is:

Among them, m represents the number of points fitting the i-th circular arc, |Q _i,j o _i | represents the distance between point Q _i,j on the i-th circular arc and the center o _i of the circle, r _i represents The radius of the i-th arc, n represents the number of arcs in a center collinear arc group, (p _x , p _y ) represents the midpoint P of the line between the two public intersection points M and N The coordinates of the point, (t _x , _ty ) represents the vector from the P point to the M point or N point, and (k ₁ ... k _n ) represents the proportional coefficient of each arc in the n arcs;

The preset nonlinear optimization algorithm is LM algorithm.

Based on the fifth feasible implementation manner of the second aspect, in the seventh feasible implementation manner of the second aspect, it is assumed that the parametric representation of any arc in any circle center collinear arc group is expressed as The coordinates of the distortion center point are (x _d0 , y _d0 ), then the distortion parameters are:

where A, B and C are the parameters in the parametric representation of any arc.

Based on the seventh feasible implementation manner of the second aspect, in the eighth feasible implementation manner of the second aspect, the distortion calibration model is:

Among them, b=(x _d , y _d ) is the coordinate point of the distorted image, a=(x _u , y _u ) is the coordinate point of the image after distortion correction, and r _d is the Euclidean distance between b and the coordinate of the distortion center point.

In the embodiment of the present invention, by acquiring a calibration image containing at least two sets of spatially parallel straight lines, the group of collinear arcs with the centers of circles mapped on the distorted image by each of the at least two sets of spatially parallel straight lines is simulated. Fit processing, and generate a distortion calibration model based on the fitting processing results. By adopting at least two sets of spatially parallel straight lines, and respectively fitting the at least two sets of spatially parallel straight lines on the center-point collinear arc group mapped on the distorted image, the calibration results of the distortion calibration model are improved. The accuracy rate; by using the collinear nature of the center of the circle, the calibration results can still maintain robustness in the presence of large noise.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only These are some embodiments of the present invention. Those skilled in the art can also obtain other drawings based on these drawings without creative work.

Fig. 1 is a schematic flowchart of a calibration method for radial distortion of a fisheye lens provided by an embodiment of the present invention;

2 is a schematic flowchart of another calibration method for radial distortion of a fisheye lens provided by an embodiment of the present invention;

3 is a schematic flowchart of another calibration method for radial distortion of a fisheye lens provided by an embodiment of the present invention;

Fig. 4 is an example schematic diagram of a group of collinear arcs with centers of circles on a distorted image provided by an embodiment of the present invention;

5 is a schematic structural diagram of a calibration device for radial distortion of a fisheye lens provided by an embodiment of the present invention;

Fig. 6 is a schematic structural diagram of an arc fitting module provided by an embodiment of the present invention;

7 is a schematic structural diagram of a parameter coefficient calculation unit provided by an embodiment of the present invention;

Fig. 8 is a schematic structural diagram of a model generation module provided by an embodiment of the present invention;

FIG. 9 is a schematic structural diagram of a coordinate parameter acquisition unit provided by an embodiment of the present invention;

FIG. 10 is a schematic structural diagram of another calibration device for radial distortion of a fisheye lens provided by an embodiment of the present invention.

detailed description

The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some, not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

The method for calibrating the radial distortion of the fisheye lens provided by the embodiment of the present invention will be described in detail below in conjunction with accompanying drawings 1 to 4 .

Please refer to FIG. 1 , which provides a schematic flowchart of a method for calibrating radial distortion of a fisheye lens according to an embodiment of the present invention. As shown in FIG. 1 , the method in the embodiment of the present invention includes the following steps S101-S103.

S101. Acquire a calibration image containing at least two sets of spatially parallel straight lines;

Specifically, the calibration device for the radial distortion of the fisheye lens acquires a calibration image taken by a fisheye lens, and the calibration image may include at least two sets of spatially parallel straight lines, and each set of spatially parallel straight lines in the at least two sets of spatially parallel straight lines Both contain several straight lines.

S102. Perform fitting processing on a group of collinear arcs with centers of circles mapped on the distorted image by each group of spatially parallel straight lines in the at least two groups of spatially parallel straight lines;

Specifically, the calibration device may operate on each set of spatially parallel straight lines in the at least two groups of spatially parallel straight lines. The calibration device performs fitting processing on each group of spatially parallel straight lines mapped on the distorted image to the center collinear arc group, and further, the calibration device uses a preset nonlinear optimization algorithm to calculate each group The common parameters and proportional coefficients of the collinear arc group with the center of the circle mapped on the distorted image by parallel straight lines in space, the calibration device uses the public parameters and the proportional coefficients to calculate and obtain fitting processing results, and the fitting processing results include A parametric representation of each arc in each of the groups of collinear arcs with centers of circles and two common intersection points corresponding to each of the groups of collinear arcs with centers of circles.

Wherein, any set of spatially parallel straight lines in the at least two groups of spatially parallel straight lines are mapped on the distorted image to a group of collinear arcs, and any one of the set of spatially parallel straight lines mapped on the distorted image There is a corresponding arc in the center collinear arc group. The public parameters include the coordinates of the midpoint of the line between the two public intersections and the vector from the midpoint to any intersection point in the two public intersections, and the proportional coefficient includes each of the collinear arcs with centers of circles The default coefficient for each arc in the group.

S103, generating a distortion calibration model according to the fitting processing result;

Specifically, the calibration device obtains the coordinates of the distortion center point and the distortion parameters according to the parameterized representation of each arc in each of the collinear arc groups and the two common intersection points corresponding to each of the collinear arc groups. , and generate a distortion calibration model by using the coordinates of the distortion center point and the distortion parameters.

In the embodiment of the present invention, by acquiring a calibration image containing at least two sets of spatially parallel straight lines, the group of collinear arcs with the centers of circles mapped on the distorted image by each of the at least two sets of spatially parallel straight lines is simulated. Fit processing, and generate a distortion calibration model based on the fitting processing results. By adopting at least two sets of spatially parallel straight lines, and respectively fitting the at least two sets of spatially parallel straight lines on the center-point collinear arc group mapped on the distorted image, the calibration results of the distortion calibration model are improved. The accuracy rate; by using the collinear nature of the center of the circle, the calibration results can still maintain robustness in the presence of large noise.

Please refer to FIG. 2 , which is a schematic flowchart of another method for calibrating radial distortion of a fisheye lens according to an embodiment of the present invention. As shown in FIG. 2 , the method in the embodiment of the present invention includes the following steps S201 to S205.

S201. Obtain a calibration image including at least two sets of spatially parallel straight lines;

S202. Using a preset nonlinear optimization algorithm, calculate the common parameters and proportional coefficients of the center collinear arc group mapped on the distorted image by each group of spatial parallel straight lines in the at least two groups of spatial parallel straight lines;

S203. Calculate and obtain a fitting processing result by using the public parameters and the proportional coefficient, the fitting processing result includes a parametric representation of each arc in each of the center-collinear arc groups and each of the center-colinear arc groups. The two public intersection points corresponding to the line arc group;

Specifically, the calibration device for the radial distortion of the fisheye lens can operate on each set of spatially parallel straight lines in the at least two groups of spatially parallel straight lines. The calibration device can use a preset nonlinear optimization algorithm to calculate the common parameters and proportional coefficients of the collinear arc group with the center of each group of parallel straight lines mapped on the distorted image. Further, the calibration device obtains the The square distance minimization equation corresponding to the center collinear arc group of each group of spatial parallel straight lines mapped on the distorted image in at least two groups of spatially parallel straight lines, and using a preset nonlinear optimization algorithm to optimize the square distance minimization equation Calculations are performed to obtain the common parameters and proportional coefficients of each of the collinear arc groups with centers.

It can be understood that each group of center-center collinear arcs has its own public parameters and proportional coefficients. The calibration device can obtain two common intersection points corresponding to each of the collinear arc groups according to the public parameters of each collinear arc group, and obtain each collinear circle according to the proportional coefficient of each collinear arc group A parametric representation of each arc in the arc group.

Wherein, any set of spatially parallel straight lines in the at least two groups of spatially parallel straight lines are mapped on the distorted image to a group of collinear arcs, and any one of the set of spatially parallel straight lines mapped on the distorted image There is a corresponding arc in the center collinear arc group. The public parameters include the coordinates of the midpoint of the line between the two public intersections and the vector from the midpoint to any intersection point in the two public intersections, and the proportional coefficient includes each of the collinear arcs with centers of circles The default coefficient for each arc in the group.

S204. Obtain the coordinates of the distortion center point and the distortion parameters according to the fitting processing result;

S205. Generate a distortion calibration model by using the coordinates of the distortion center point and the distortion parameters;

Specifically, the calibration device acquires the straight line equations of the two common intersection points corresponding to the collinear arc groups with the centers of each circle, and determines the coordinates of the intersection points of all the straight line equations as the coordinates of the distortion center point. Coordinates and the parametric representation of any arc in the group of arcs whose centers are collinear, get the distortion parameters. The calibration device may generate a distortion calibration model according to the distortion midpoint coordinates and the distortion parameters.

Wherein, for step S201 in the embodiment of the present invention, reference may be made to step S101 in the embodiment shown in FIG. 1 , and details are not described here.

In the embodiment of the present invention, by acquiring a calibration image containing at least two sets of spatially parallel straight lines, the group of collinear arcs with the centers of circles mapped on the distorted image by each of the at least two sets of spatially parallel straight lines is simulated. Fit processing, and generate a distortion calibration model based on the fitting processing results. By adopting at least two sets of spatially parallel straight lines, and respectively fitting the at least two sets of spatially parallel straight lines on the center-point collinear arc group mapped on the distorted image, the calibration results of the distortion calibration model are improved. The accuracy rate; by using the collinear nature of the center of the circle, the calibration results can still maintain robustness in the presence of large noise.

Please refer to FIG. 3 , which is a schematic flowchart of another method for calibrating radial distortion of a fisheye lens according to an embodiment of the present invention. As shown in FIG. 3 , the method in the embodiment of the present invention may include the following steps S301-S307.

S301. Acquire a calibration image including at least two sets of spatially parallel straight lines;

Specifically, the calibration device for the radial distortion of the fisheye lens acquires a calibration image taken by a fisheye lens, and the calibration image may include at least two sets of spatially parallel straight lines, and each set of spatially parallel straight lines in the at least two sets of spatially parallel straight lines Both contain several straight lines.

In the embodiment of the present invention, the adopted distortion calibration model can be expressed as formula (1):

In formula (1), b=(x _d , y _d ) is the coordinate point of the distorted image, a=(x _u , y _u ) is the coordinate point of the image after distortion correction, and r _d is the relationship between b and the coordinates of the distortion center point Euclidean distance. Since the traditional radial distortion distortion calibration model is based on the assumption that the distortion center point is the center of the image, the distortion parameter λ is solved to generate the distortion calibration model, but in general, the distortion center point is not located at the image center, Therefore, in the embodiment of the present invention, the coordinates (x _d0 , y _d0 ) of the distortion center point are also used as unknown parameters for solution.

In three-dimensional space, the equation of any line in each group of parallel lines in space can be expressed as formula (2):

In formula (2), D=(dx, dy, dz) is the unit direction vector, ( _xi , y _i , _zi ) is a certain three-dimensional space point that the straight line passes through, and t is a scale factor used to generate different points on this line. It can be known from formula (2) that a group of parallel straight lines in space may pass through different points ( _xi , y, _{zi )} , but must have the same unit direction vector D.

Under ideal perspective projection, the group of spatially parallel straight lines is mapped to a group of straight lines on the ideal perspective image, which can be expressed as formula (3):

(3) where f is the focal length of the ideal perspective camera. When the passing three-dimensional points ( _xi , y, _{zi )} are different, the lines formed by the projection of the set of parallel straight lines in space onto the ideal perspective image may no longer be parallel.

When t→±∞, the projection point of the vanishing point of this group of parallel lines in space on the ideal perspective image can be obtained, which can be expressed as formula (4):

It can be seen from formula (4) that under the ideal perspective projection, the set of parallel spatial lines will intersect at the same point on the ideal perspective image (that is, the projection point of the vanishing point of the parallel spatial lines on the ideal perspective image).

Substituting formula (1) into formula (4), and assuming that the distortion center is the origin of the image, formula (5) can be obtained:

And (6) type:

Divide formula (5) by formula (6) to get:

Equation (7) indicates that under the traditional radial distortion distortion calibration model, a and b have the same slope, and they are collinear with the three points of the image coordinate origin. Substituting formula (7) into formula (6), formula (8) can be obtained:

This is a one-variable quadratic equation about y _d , and its discriminant formula is (9):

As for the radial distortion, λ<0, and thus Δ must be greater than zero, so it can be obtained that formula (8) must have 2 real number solutions. From the above derivation, it can be seen that under the traditional radial distortion distortion calibration model, the vanishing points of a group of parallel straight lines in space will have two projection points on the distorted image. Since a group of parallel straight lines in space is mapped into a center-collinear arc group under this model, all arcs in the center-collinear arc group will have two common intersection points.

It should be noted that any set of spatially parallel straight lines in the at least two groups of spatially parallel straight lines are mapped to a group of collinear arcs with the centers of circles on the distorted image, and any one of the set of spatially parallel straight lines in the distorted image There is a corresponding arc in the center collinear arc group mapped above.

S302. Acquire a square distance minimization equation corresponding to a group of collinear arcs with center points mapped on the distorted image by each group of spatially parallel straight lines in the at least two groups of spatially parallel straight lines;

Specifically, when the calibration device fits the group of collinear arcs with the center of the circle mapped on the distorted image by each of the at least two sets of parallel straight lines in space, the calibration device obtains the collinear arc with the center of the circle The group's squared distance minimization equation.

Please also refer to FIG. 4 , which provides an example schematic diagram of a group of collinear arcs with centers of circles on a distorted image according to an embodiment of the present invention. As shown in Figure 4, assuming that the two common intersection points are points M and N respectively, assuming that the coordinates of the midpoint P on the line connecting M and N points on the distorted image are (p _x , p _y ), obtain points P to N The vector of point or M point, adopt the vector of P point to N point in the embodiment of the present invention, denote as If the center position of the i-th arc in the collinear arc group is o _i , then the unit vector of the directed line from point P to point o _i can be expressed as formula (10):

(10) where, representation and vector Perpendicular vectors must satisfy representation vector model. It can be seen that Po _i ^o has two sets of solutions, which are with

set a scaling factor Then the center o _i can be expressed as (11):

Assuming that the collinear circle group corresponding to the center collinear arc group has n arcs, the total number of unknown variables to be obtained is (4+n), and n in (4+n) means that there are n proportional coefficients (k ₁ ... k _n ), 4 represents the public parameters shared by n arcs, which are respectively p _x and p _y representing the position of point P, and representing the vector t _x and t _y .

For the i-th arc in the collinear circle group, its radius can be expressed as formula (12):

Then when fitting the arc, the equation required to be minimized is (13):

In formula (13), m represents the number of points fitting the arc, and |Q _i,j o _i | is the distance between point Q _i,j on the i-th arc and the center o _i .

In a group of collinear arcs with centers, there are n such arcs, so the square distance minimization equation is (14):

S303. Using a preset nonlinear optimization algorithm to calculate the square distance minimization equation, and obtain the common parameters and proportional coefficients of each of the collinear arc groups with centers of circles;

Specifically, the preset nonlinear optimization algorithm is preferably the LM algorithm, and the calibration device uses the preset nonlinear optimization algorithm to calculate the square distance minimization equation to obtain the collinear circles with the centers of each Common parameters and scale factor for the arc group.

Wherein, the public parameters include the midpoint coordinates of the line between the two public intersections and the vector from the midpoint to any intersection point of the two public intersections, and the proportional coefficient includes the collinear The preset coefficient corresponding to each arc in the arc group.

S304. Calculate and obtain a fitting processing result by using the public parameter and the proportional coefficient, and the fitting processing result includes a parameterized representation of each arc in each of the center-collinear arc groups and each of the center-colinear arc groups. The two public intersection points corresponding to the line arc group;

Specifically, each center-center collinear arc group has its own public parameters and proportional coefficients. The calibration device can obtain two common intersection points corresponding to each of the collinear arc groups according to the public parameters of each collinear arc group, and obtain each collinear circle according to the proportional coefficient of each collinear arc group A parametric representation of each arc in the arc group.

S305. Obtain the straight line equations of the two common intersection points corresponding to each circle center collinear arc group, and determine the intersection point coordinates of all the straight line equations as the distortion center point coordinates;

Specifically, the calibration device obtains the straight line equation formed by the two public intersections corresponding to the collinear arc groups with the centers of each circle, for example: the straight line equation determined by point M and point N in the distorted image, if there are parallel straight lines in space W group, then there are also W corresponding circle center collinear arc groups, then W straight line equations can be obtained, the intersection point coordinates are calculated for the W straight line equations, and the intersection point coordinates are determined as the distortion center point coordinates. For example: the coordinates of the intersection obtained from the equations of W straight lines are (x _d0 , y _d0 ), then (x _d0 , y _d0 ) is determined as the coordinates of the distortion center point.

S306. Obtain a distortion parameter according to the coordinates of the distortion center point and the parametric representation of any arc in any center collinear arc group;

Specifically, it is assumed that the parametric expression of any arc in any of the obtained collinear arc groups is as The coordinates of the distortion center point are (x _d0 , y _d0 ), then the distortion parameters are:

where A, B and C are the parameters in the parametric representation of any arc.

S307. Generate a distortion calibration model by using the coordinates of the distortion center point and the distortion parameters;

Specifically, the calibration device can determine the expression of r _d according to the coordinates of the distortion center point. Assuming that the coordinate of the distortion center point is c, the expression method is:

r _d =︱bc︱

Substitute the expression and distortion parameter λ into formula (1) to generate a distortion calibration model.

In the embodiment of the present invention, by acquiring a calibration image containing at least two sets of spatially parallel straight lines, the group of collinear arcs with the centers of circles mapped on the distorted image by each of the at least two sets of spatially parallel straight lines is simulated. Fit processing, and generate a distortion calibration model based on the fitting processing results. By adopting at least two sets of spatially parallel straight lines, respectively performing fitting processing on the center collinear arc groups mapped on the distorted image by each set of spatially parallel straight lines in the at least two sets of spatially parallel straight lines, and recalculating the coordinates of the distortion center point , which improves the accuracy of the calibration results of the distortion calibration model; by using the collinear property of the center of the circle, the calibration results can still maintain robustness in the presence of large noise.

The apparatus for calibrating the radial distortion of the fisheye lens provided by the embodiment of the present invention will be described in detail below with reference to FIG. 5 and FIG. 9 . It should be noted that the calibration device shown in Figure 5 and Figure 9 is used to execute the method of the embodiment shown in Figure 1-Figure 4 of the present invention, for the convenience of description, only shows the For the part, the specific technical details are not disclosed, please refer to the embodiment shown in Fig. 1-Fig. 4 of the present invention.

Please refer to FIG. 5 , which provides a schematic structural diagram of an apparatus for calibrating radial distortion of a fisheye lens according to an embodiment of the present invention. As shown in FIG. 5 , the calibration device 1 of the embodiment of the present invention may include: a calibration image acquisition module 11 , an arc fitting module 12 and a model generation module 13 .

A calibration image acquisition module 11, configured to acquire a calibration image comprising at least two groups of spatially parallel straight lines;

In a specific implementation, the calibration image acquisition module 11 acquires a calibration image captured by a fisheye lens, and the calibration image may include at least two sets of spatially parallel straight lines, and each set of spatially parallel straight lines in the at least two sets of spatially parallel straight lines is Contains several straight lines.

In the embodiment of the present invention, the adopted distortion calibration model can be expressed as formula (1):

In formula (1), b=(x _d , y _d ) is the coordinate point of the distorted image, a=(x _u , y _u ) is the coordinate point of the image after distortion correction, and r _d is the relationship between b and the coordinates of the distortion center point Euclidean distance. Since the traditional radial distortion distortion calibration model is based on the assumption that the distortion center point is the center of the image, the distortion parameter λ is solved to generate the distortion calibration model, but in general, the distortion center point is not located at the image center, Therefore, in the embodiment of the present invention, the coordinates (x _d0 , y _d0 ) of the distortion center point are also used as unknown parameters for solution.

In three-dimensional space, the equation of any line in each group of parallel lines in space can be expressed as formula (2):

In formula (2), D=(dx, dy, dz) is the unit direction vector, ( _xi , y _i , _zi ) is a certain three-dimensional space point that the straight line passes through, and t is a scale factor used to generate different points on this line. It can be known from formula (2) that a group of parallel straight lines in space may pass through different points ( _xi , y, _{zi )} , but must have the same unit direction vector D.

Under ideal perspective projection, the group of spatially parallel straight lines is mapped to a group of straight lines on the ideal perspective image, which can be expressed as formula (3):

(3) where f is the focal length of the ideal perspective camera. When the passing three-dimensional points ( _xi , y, _{zi )} are different, the lines formed by the projection of the set of parallel straight lines in space onto the ideal perspective image may no longer be parallel.

When t→±∞, the projection point of the vanishing point of this group of parallel lines in space on the ideal perspective image can be obtained, which can be expressed as formula (4):

It can be seen from formula (4) that under the ideal perspective projection, the set of parallel spatial lines will intersect at the same point on the ideal perspective image (that is, the projection point of the vanishing point of the parallel spatial lines on the ideal perspective image).

Substituting formula (1) into formula (4), and assuming that the distortion center is the origin of the image, formula (5) can be obtained:

And (6) type:

Divide formula (5) by formula (6) to get:

Equation (7) indicates that under the traditional radial distortion distortion calibration model, a and b have the same slope, and they are collinear with the three points of the image coordinate origin. Substituting formula (7) into formula (6), formula (8) can be obtained:

This is a one-variable quadratic equation about y _d , and its discriminant formula is (9):

As for the radial distortion, λ<0, Δ must be greater than zero, so it can be obtained that formula (8) must have 2 real number solutions. From the above derivation, it can be seen that under the traditional radial distortion distortion calibration model, the vanishing points of a group of parallel straight lines in space will have two projection points on the distorted image. Since a group of parallel straight lines in space is mapped into a center-collinear arc group under this model, all arcs in the center-collinear arc group will have two common intersection points.

It should be noted that any set of spatially parallel straight lines in the at least two groups of spatially parallel straight lines are mapped to a group of collinear arcs with the centers of circles on the distorted image, and any one of the set of spatially parallel straight lines in the distorted image There is a corresponding arc in the center collinear arc group mapped above.

An arc fitting module 12, configured to perform fitting processing on the group of collinear arcs with the center of the circle mapped on the distorted image by each group of spatial parallel straight lines in the at least two groups of spatially parallel straight lines;

In a specific implementation, the arc fitting module 12 may separately operate on each group of spatially parallel straight lines in the at least two groups of spatially parallel straight lines. The arc fitting module 12 performs fitting processing on each group of spatially parallel straight lines mapped on the distorted image to the center collinear arc group, further, the arc fitting module 12 adopts a preset nonlinear optimization Algorithm, calculating the common parameters and proportional coefficients of the center collinear arc group mapped on the distorted image by each group of spatially parallel straight lines, the arc fitting module 12 uses the public parameters and the proportional coefficient to calculate and obtain A fitting processing result, the fitting processing result including a parametric representation of each arc in each of the groups of collinear arcs with centers of circles and two common intersection points corresponding to each of the groups of collinear arcs with centers of circles.

Wherein, the public parameters include the midpoint coordinates of the line between the two public intersections and the vector from the midpoint to any intersection point of the two public intersections, and the proportional coefficient includes the collinear The preset coefficient corresponding to each arc in the arc group.

Specifically, please refer to FIG. 6 , which provides a schematic structural diagram of an arc fitting module according to an embodiment of the present invention. As shown in Figure 6, the arc fitting module 12 may include:

The parameter coefficient calculation unit 121 is configured to use a preset nonlinear optimization algorithm to calculate the common parameters and proportions of the center collinear arc group mapped on the distorted image by each set of spatial parallel straight lines in the at least two groups of spatial parallel straight lines coefficient;

The result obtaining unit 122 is configured to use the public parameters and the proportional coefficient to calculate and obtain a fitting processing result, the fitting processing result including a parametric representation of each arc in each of the collinear arc groups with centers of circles and Two public intersection points corresponding to each of the collinear arc groups with centers of circles;

In a specific implementation, the parameter coefficient calculating unit 121 may separately operate on each group of spatially parallel straight lines in the at least two groups of spatially parallel straight lines. The parameter coefficient calculation unit 121 can use a preset nonlinear optimization algorithm to calculate the public parameters and proportional coefficients of the center collinear arc group mapped by each group of spatially parallel straight lines on the distorted image. Further, the parameter coefficient The calculation unit 121 obtains the square distance minimization equation corresponding to the collinear arc group with the center of the center of each group of the at least two groups of spatially parallel straight lines mapped on the distorted image, and uses a preset nonlinear optimization algorithm to The square distance minimization equation is used for calculation, and the common parameters and proportional coefficients of each of the collinear arc groups with the centers of the circles are obtained.

It can be understood that each group of center-center collinear arcs has its own public parameters and proportional coefficients. The result acquisition unit 122 can obtain the two common intersection points corresponding to each of the collinear arc groups according to the common parameters of each group of collinear arcs, and obtain the common points of intersection of each group of collinear arcs according to the proportional coefficient of each group of collinear arcs. A parametric representation of each arc in the line-arc group.

Wherein, the public parameters include the midpoint coordinates of the line between the two public intersections and the vector from the midpoint to any intersection point of the two public intersections, and the proportional coefficient includes the collinear The preset coefficient corresponding to each arc in the arc group.

Specifically, please refer to FIG. 7 , which provides a schematic structural diagram of a parameter coefficient calculation unit for an embodiment of the present invention. As shown in Figure 7, the parameter coefficient calculation unit 121 may include:

The equation obtaining subunit 1211 is configured to obtain the square distance minimization equation corresponding to the group of collinear arcs with the center of the circle mapped on the distorted image by each group of the at least two groups of spatially parallel straight lines;

In a specific implementation, when the equation acquisition subunit 1211 fits the group of collinear arcs with the centers of the circles mapped on the distorted image by each set of spatial parallel straight lines in the at least two sets of spatial parallel straight lines, the equation acquisition subunit 1211 Obtains the squared distance minimization equation for this set of collinear arcs with centers.

Please also refer to FIG. 4 , which provides an example schematic diagram of a group of collinear arcs with centers of circles on a distorted image according to an embodiment of the present invention. As shown in Figure 4, assuming that the two common intersection points are points M and N respectively, assuming that the coordinates of the midpoint P on the line connecting M and N points on the distorted image are (p _x , p _y ), obtain points P to N The vector of point or M point, adopt the vector of P point to N point in the embodiment of the present invention, denote as If the center position of the i-th arc in the collinear arc group is o _i , then the unit vector of the directed line from point P to point o _i can be expressed as formula (10):

(10) where, representation and vector Perpendicular vectors must satisfy representation vector model. It can be seen that Po _i ^o has two sets of solutions, which are with

set a scaling factor Then the center o _i can be expressed as (11):

Assuming that the collinear circle group corresponding to the center collinear arc group has n arcs, the total number of unknown variables to be obtained is (4+n), and n in (4+n) means that there are n proportional coefficients (k ₁ ... k _n ), 4 represents the public parameters shared by n arcs, which are respectively p _x and p _y representing the position of point P, and representing the vector t _x and t _y .

For the i-th arc in the collinear circle group, its radius can be expressed as formula (12):

Then when fitting the arc, the equation required to be minimized is (13):

In formula (13), m represents the number of points fitting the arc, and |Q _i,j o _i | is the distance between point Q _i,j on the i-th arc and the center o _i .

In a group of collinear arcs with centers, there are n such arcs, so the square distance minimization equation is (14):

The parameter coefficient obtaining subunit 1212 is used to calculate the square distance minimization equation by using a preset nonlinear optimization algorithm, and obtain the public parameters and proportional coefficients of each of the collinear arc groups with centers of circles;

In a specific implementation, the preset nonlinear optimization algorithm is preferably the LM algorithm, and the parameter coefficient acquisition subunit 1212 uses the preset nonlinear optimization algorithm to calculate the square distance minimization equation to obtain each Describe the common parameters and proportional coefficients of the center collinear arc group.

Wherein, the public parameters include the midpoint coordinates of the line between the two public intersections and the vector from the midpoint to any intersection point of the two public intersections, and the proportional coefficient includes the collinear The preset coefficient corresponding to each arc in the arc group.

A model generation module 13, configured to generate a distortion calibration model according to the fitting processing result;

In a specific implementation, the model generation module 13 obtains the coordinates of the distortion center point according to the parameterized representation of each arc in each of the collinear arc groups and the two public intersection points corresponding to each of the collinear arc groups. and distortion parameters, and generate a distortion calibration model by using the coordinates of the distortion center point and the distortion parameters.

Specifically, please refer to FIG. 8 , which provides a schematic structural diagram of a model generation module for an embodiment of the present invention. As shown in Figure 8, the model generation module 13 may include:

A coordinate parameter acquiring unit 131, configured to acquire the coordinates of the distortion center point and the distortion parameters according to the fitting processing result;

In a specific implementation, the coordinate parameter acquisition unit 131 acquires the straight line equations of the two common intersection points corresponding to each circle center collinear arc group, and determines the intersection point coordinates of all the straight line equations as the distortion center point coordinates, and the coordinate parameter acquisition unit 131 Obtain a distortion parameter according to the coordinates of the distortion center point and the parametric representation of any arc in any circle center collinear arc group.

Specifically, please refer to FIG. 9 , which provides a schematic structural diagram of a coordinate parameter acquisition unit for an embodiment of the present invention. As shown in Figure 9, the coordinate parameter acquisition unit 131 may include:

The coordinate obtaining subunit 1311 is used to obtain the straight line equations of the two common intersection points corresponding to the collinear arc groups of the respective circle centers, and determine the intersection point coordinates of all the straight line equations as the distortion center point coordinates;

In a specific implementation, the coordinate acquisition subunit 1311 acquires the straight line equation formed by the two common intersection points corresponding to the collinear arc groups with the centers of each circle, for example: the straight line equation determined by point M and point N in the distorted image, if the space If there are W groups of parallel straight lines, there are also W corresponding circle center collinear arc groups, then W straight line equations can be obtained, the coordinates of the intersection points are calculated for the W straight line equations, and the coordinates of the intersection points are determined as the coordinates of the distortion center point. For example: the coordinates of the intersection obtained from the equations of W straight lines are (x _d0 , y _d0 ), then (x _d0 , y _d0 ) is determined as the coordinates of the distortion center point.

A parameter acquisition subunit 1312, configured to acquire distortion parameters according to the coordinates of the distortion center point and the parametric representation of any arc in any circle center collinear arc group;

In the specific implementation, it is assumed that the parameterized expression of any arc in any circle center collinear arc group obtained by the parameter acquisition subunit 1312 is expressed as The coordinates of the distortion center point are (x _d0 , y _d0 ), then the distortion parameters are:

where A, B and C are the parameters in the parametric representation of any arc.

A model generating unit 132, configured to generate a distortion calibration model by using the coordinates of the distortion center point and the distortion parameters;

In a specific implementation, the model generation unit 132 may generate a distortion calibration model according to the distortion midpoint coordinates and the distortion parameters.

The model generation unit 132 can determine the expression of r according to the coordinates of the distortion center point. Assuming that the coordinate of the distortion center point is _c , the expression is:

r _d =︱bc︱

The model generation unit 132 substitutes the expression and the distortion parameter λ into formula (1) to generate a distortion calibration model.

In the embodiment of the present invention, by acquiring a calibration image containing at least two sets of spatially parallel straight lines, the group of collinear arcs with the centers of circles mapped on the distorted image by each of the at least two sets of spatially parallel straight lines is simulated. Fit processing, and generate a distortion calibration model based on the fitting processing results. By adopting at least two sets of spatially parallel straight lines, respectively performing fitting processing on the center collinear arc groups mapped on the distorted image by each set of spatially parallel straight lines in the at least two sets of spatially parallel straight lines, and recalculating the coordinates of the distortion center point , which improves the accuracy of the calibration results of the distortion calibration model; by using the collinear property of the center of the circle, the calibration results can still maintain robustness in the presence of large noise.

Please refer to FIG. 10 , which provides a schematic structural diagram of another calibration device for radial distortion of a fisheye lens according to an embodiment of the present invention. As shown in FIG. 10 , the calibration device 1000 may include: at least one processor 1001 , such as a CPU, a user interface 1003 , at least one network interface 1004 , a memory 1005 , and at least one communication bus 1002 . Wherein, the communication bus 1002 is used to realize connection and communication between these components. Wherein, the user interface 1003 may include a display screen (Display) and a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. Optionally, the network interface 1004 may include a standard wired interface and a wireless interface (such as a WI-FI interface). The memory 1005 may be a high-speed RAM memory, or a non-volatile memory, such as at least one disk memory. Optionally, the memory 1005 may also be at least one storage device located away from the aforementioned processor 1001 . As shown in FIG. 10 , the memory 1005 as a computer storage medium may include an operating system, a network communication module, a user interface module, and a calibration application program.

The processor 1001 can be used to call the calibration application program stored in the memory 1005, and specifically perform the following steps:

Obtain a calibration image comprising at least two sets of spatially parallel straight lines;

performing fitting processing on the group of collinear arcs with the center of the circle mapped on the distorted image by each of the at least two groups of spatially parallel straight lines;

A distortion calibration model is generated according to the fitting processing results.

In a feasible implementation manner of the embodiment of the present invention, any set of spatially parallel straight lines among the at least two groups of spatially parallel straight lines are mapped to a group of collinear arcs with centers of circles on the distorted image, and any set of spatially parallel straight lines Any straight line among the straight lines has a corresponding circular arc in the group of collinear arcs with the center of the circle mapped on the distorted image.

In another feasible implementation manner of the embodiment of the present invention, the processor 1001 executes the group of collinear arcs with the centers of the circles mapped on the distorted image for each set of the at least two sets of spatially parallel straight lines. When performing fitting processing, the following steps are specifically performed:

Using a preset non-linear optimization algorithm to calculate the common parameters and proportional coefficients of the center collinear arc group mapped on the distorted image by each group of spatial parallel straight lines in the at least two groups of spatial parallel straight lines;

Using the public parameters and the proportional coefficient to calculate and obtain a fitting processing result, the fitting processing result includes a parametric representation of each arc in each of the center collinear arc groups and each of the center collinear circles The two public intersection points corresponding to the arc group;

Wherein, the public parameters include the midpoint coordinates of the line between the two public intersections and the vector from the midpoint to any intersection point of the two public intersections, and the proportional coefficient includes the collinear The preset coefficient corresponding to each arc in the arc group.

In yet another feasible implementation manner of the embodiment of the present invention, the processor 1001 is executing a preset nonlinear optimization algorithm to calculate the effect of each group of spatially parallel straight lines in the at least two groups of spatially parallel straight lines on the distorted image. When the public parameters and proportional coefficients of the center-point collinear arc group are mapped, the following steps are specifically performed:

Obtaining the square distance minimization equation corresponding to the group of collinear arcs with the centers of each group of the at least two groups of spatially parallel straight lines mapped on the distorted image;

A preset nonlinear optimization algorithm is used to calculate the square distance minimization equation to obtain the common parameters and proportional coefficients of each of the collinear arc groups with centers.

In yet another feasible implementation manner of the embodiment of the present invention, when the processor 1001 generates the distortion calibration model according to the fitting processing result, specifically perform the following steps:

Obtain the coordinates of the distortion center point and the distortion parameters according to the fitting processing result;

A distortion calibration model is generated by using the coordinates of the distortion center point and the distortion parameters.

In yet another feasible implementation manner of the embodiment of the present invention, when the processor 1001 executes obtaining the coordinates of the distortion center point and the distortion parameters according to the fitting processing result, specifically perform the following steps:

Obtain the straight line equations of the two common intersection points corresponding to the collinear arc groups of the respective circle centers, and determine the intersection coordinates of all straight line equations as the distortion center point coordinates;

The distortion parameters are obtained according to the coordinates of the distortion center point and the parametric representation of any circular arc in any circle center collinear circular arc group.

In yet another feasible implementation manner of the embodiment of the present invention, the square distance minimization equation is:

Among them, m represents the number of points fitting the i-th circular arc, |Q _i,j o _i | represents the distance between point Q _i,j on the i-th circular arc and the center o _i of the circle, r _i represents The radius of the i-th arc, n represents the number of arcs in a center collinear arc group, (p _x , p _y ) represents the midpoint P of the line between the two public intersection points M and N The coordinates of the point, (t _x , _ty ) represents the vector from the P point to the M point or N point, and (k ₁ ... k _n ) represents the proportional coefficient of each arc in the n arcs;

The preset nonlinear optimization algorithm is LM algorithm.

In yet another feasible implementation of the embodiment of the present invention, it is assumed that the parametric representation of any arc in any group of collinear arcs is The coordinates of the distortion center point are (x _d0 , y _d0 ), then the distortion parameters are:

where A, B and C are the parameters in the parametric representation of any arc.

In yet another feasible implementation manner of the embodiment of the present invention, the distortion calibration model is:

Among them, b=(x _d , y _d ) is the coordinate point of the distorted image, a=(x _u , y _u ) is the coordinate point of the image after distortion correction, and r _d is the Euclidean distance between b and the coordinate of the distortion center point.

In the embodiment of the present invention, by acquiring a calibration image containing at least two sets of spatially parallel straight lines, the group of collinear arcs with the centers of circles mapped on the distorted image by each of the at least two sets of spatially parallel straight lines is simulated. Fit processing, and generate a distortion calibration model based on the fitting processing results. By adopting at least two sets of spatially parallel straight lines, respectively performing fitting processing on the center collinear arc groups mapped on the distorted image by each set of spatially parallel straight lines in the at least two sets of spatially parallel straight lines, and recalculating the coordinates of the distortion center point , which improves the accuracy of the calibration results of the distortion calibration model; by using the collinear property of the center of the circle, the calibration results can still maintain robustness in the presence of large noise.

Through the above description of the implementation manners, those skilled in the art can clearly understand that the present invention can be implemented by hardware, firmware, or a combination thereof. When implemented in software, the functions described above may be stored on or transmitted over as one or more instructions or code on a computer-readable medium. Computer-readable media includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one place to another. A storage media may be any available media that can be accessed by a computer. By way of example and without limitation, computer readable media may include RAM, ROM, or other optical disk storage, magnetic disk storage media, or other magnetic storage devices, or may be used to carry or store desired program code in the form of instructions or data structures and any other medium that can be accessed by a computer. also. Any connection can suitably be a computer-readable medium. For example, if the software is transmitted from a website, server, or other remote source using coaxial cable, fiber optic cable, twisted pair, digital subscriber line, or wireless technologies such as infrared, radio, and microwave, then the coaxial cable, fiber optic cable , twisted pair, digital subscriber line, or wireless technologies such as infrared, wireless, and microwave are included in the fixation of the respective media. As used herein, disk and disc include compact disc, laser disc, optical disc, digital versatile disc, floppy disc and blu-ray disc, where discs usually reproduce data magnetically, while discs reproduce data optically with lasers Copy data. Combinations of the above should also be included within the scope of computer-readable media.

The above disclosures are only preferred embodiments of the present invention, and certainly cannot limit the scope of rights of the present invention. Therefore, equivalent changes made according to the claims of the present invention still fall within the scope of the present invention.

Claims (16)

1. a kind of scaling method of fish eye lens radial distortion, it is characterised in that include：

Obtain the uncalibrated image comprising at least two groups of space parallel lines；

The center of circle mapped on fault image each group of space parallel lines at least two groups space parallel lines is conllinear Circular arc group is fitted process；

Distortion peg model is generated according to process of fitting treatment result；

Wherein, it is described that each group of space parallel lines at least two groups space parallel lines mapped on fault image The conllinear circular arc group in the center of circle is fitted process, including：

Using default nonlinear optimization algorithm, each group of space parallel lines at least two groups space parallel lines described in calculating The common parameter and proportionality coefficient of the conllinear circular arc group in the center of circle mapped on fault image；

Calculated using the common parameter and the proportionality coefficient and obtain process of fitting treatment result, the process of fitting treatment result includes each The parametrization of every circular arc is represented and the conllinear circular arc group corresponding two in each described center of circle in the conllinear circular arc group in the individual center of circle Individual common intersection；

Wherein, the common parameter includes that the middle point coordinates of line and the midpoint are to described two between described two common intersections The vector of arbitrary intersection point in individual common intersection, the proportionality coefficient includes every circular arc pair in the conllinear circular arc group in each described center of circle The predetermined coefficient answered.

2. method according to claim 1, it is characterised in that arbitrary group of space at least two groups space parallel lines Parallel lines is mapped as the conllinear circular arc group in center of circle on fault image, arbitrary straight in the arbitrary group of space parallel lines There is a corresponding circular arc in the conllinear circular arc group in the center of circle that line maps on fault image.

3. method according to claim 1, it is characterised in that described using default nonlinear optimization algorithm, calculates institute State at least two groups space parallel lines the conllinear circular arc group in the center of circle that each group of space parallel lines maps on fault image Common parameter and proportionality coefficient, including：

Obtain at least two groups space parallel lines center of circle that each group of space parallel lines map on fault image common The corresponding square distance minimization equation of line circular arc group；

The square distance minimization equation is calculated using default nonlinear optimization algorithm, obtains each center of circle The common parameter and proportionality coefficient of conllinear circular arc group.

4. method according to claim 1, it is characterised in that described that distortion calibration mold is generated according to process of fitting treatment result Type, including：

Center of distortion point coordinates and distortion parameter are obtained according to process of fitting treatment result；

Distortion peg model is generated using the center of distortion point coordinates and the distortion parameter.

5. method according to claim 4, it is characterised in that described center of distortion point is obtained according to process of fitting treatment result to sit Mark and distortion parameter, including：

Obtain the linear equation of corresponding two common intersections of the conllinear circular arc group in each center of circle, and by the intersection point of all linear equations Coordinate is defined as center of distortion point coordinates；

Represented according to the parametrization of arbitrary circular arc in the conllinear circular arc group of the center of distortion point coordinates and arbitrary center of circle, obtain distortion Parameter.

6. method according to claim 3, it is characterised in that the square distance minimization equation is：

$\underset{\left\{\begin{array}{c}{p}_x,p_y,t_x,t_y\\ k_1\mathrm{...}{k}_n\end{array}\right\}}{\arg \min }\underset{i=1}{\overset{n}{\&Sigma;}}\underset{j=1}{\overset{m}{\&Sigma;}}{\&lsqb;\left|Q_{i,j}{o}_i\right|-r_i\&rsqb;}^2$

Wherein, m represents the number of the point of i-th circular arc of fitting, | Q_i,jo_i| represent the point Q on i-th circular arc_i,jWith center of circle o_iIt Between distance, r_iThe radius of i-th circular arc is represented, n represents the bar number of circular arc in the conllinear circular arc group in center of circle, (p_x, p_y) represent The coordinate of the midpoint P points of line, (t between described two common intersection M points and N points_x, t_y) represent the P points to the M points or N points Vector, (k₁…k_n) represent n bar circular arcs in every circular arc proportionality coefficient；

The default nonlinear optimization algorithm is LM (Leven-Marquardt) algorithm.

7. method according to claim 5, it is characterised in that assume the ginseng of arbitrary circular arc in the conllinear circular arc group in arbitrary center of circle Numberization is expressed asThe center of distortion point coordinates is (x_d0, y_d0), then the distortion parameter For：

$\mathrm{\&lambda;}=\frac{1}{x_{d0}^2+y_{d0}^2+{\mathrm{Ax}}_{d0}+{\mathrm{By}}_{d0}+C}$

Wherein, A, B and C are the parameter in the parametrization expression of arbitrary circular arc.

8. method according to claim 7, it is characterised in that the distortion peg model is：

$a=\frac{b}{1+{\mathrm{\&lambda;r}}_d^2}$

Wherein, b=(x_d, y_d) be fault image coordinate points, a=(x_u, y_u) it is image coordinate point after distortion correction, r_dFor The Euclidean distance of b and center of distortion point coordinates.

9. a kind of caliberating device of fish eye lens radial distortion, it is characterised in that include：

Uncalibrated image acquisition module, for obtaining the uncalibrated image comprising at least two groups of space parallel lines；

Circular fitting module, for each group of space parallel lines at least two groups space parallel lines in fault image The conllinear circular arc group in the center of circle of upper mapping is fitted process；

Model generation module, for generating distortion peg model according to process of fitting treatment result；

Wherein, the circular fitting module includes：

Parameter coefficient calculation unit, for using default nonlinear optimization algorithm, at least two groups space straight parallels described in calculating The common parameter and proportionality coefficient of the conllinear circular arc group in the center of circle that each group of space parallel lines maps on fault image in line；

As a result acquiring unit, for calculating using the common parameter and the proportionality coefficient process of fitting treatment result is obtained, described Process of fitting treatment result includes that the parametrization of every circular arc in the conllinear circular arc group in each described center of circle is represented and each described center of circle Corresponding two common intersections of conllinear circular arc group；

Wherein, the common parameter includes that the middle point coordinates of line and the midpoint are to described two between described two common intersections The vector of arbitrary intersection point in individual common intersection, the proportionality coefficient includes every circular arc pair in the conllinear circular arc group in each described center of circle The predetermined coefficient answered.

10. caliberating device according to claim 9, it is characterised in that arbitrary at least two groups space parallel lines Group space parallel lines is mapped as the conllinear circular arc group in center of circle on fault image, in the arbitrary group of space parallel lines There is a corresponding circular arc in the conllinear circular arc group in the center of circle that arbitrary straight line maps on fault image.

11. caliberating devices according to claim 9, it is characterised in that the parameter coefficient calculation unit includes：

Equation obtains subelement, for obtaining at least two groups space parallel lines in each group of space parallel lines in distortion The corresponding square distance minimization equation of the conllinear circular arc group in the center of circle mapped on image；

Parameter coefficient obtains subelement, for being entered to the square distance minimization equation using default nonlinear optimization algorithm Row is calculated, and obtains the common parameter and proportionality coefficient of the conllinear circular arc group in each described center of circle.

12. caliberating devices according to claim 9, it is characterised in that the model generation module includes：

Coordinate parameters acquiring unit, for obtaining center of distortion point coordinates and distortion parameter according to process of fitting treatment result；

Model signal generating unit, for generating distortion peg model using the center of distortion point coordinates and the distortion parameter.

13. caliberating devices according to claim 12, it is characterised in that the coordinate parameters acquiring unit includes：

Coordinate obtains subelement, for obtaining the linear equation of corresponding two common intersections of the conllinear circular arc group in each center of circle, and The intersecting point coordinate of all linear equations is defined as into center of distortion point coordinates；

Parameter acquiring subelement, for according to arbitrary circular arc in the conllinear circular arc group of the center of distortion point coordinates and arbitrary center of circle Parametrization expression, obtains distortion parameter.

14. caliberating devices according to claim 11, it is characterised in that the square distance minimization equation is：

$\underset{\left\{\begin{array}{c}{p}_x,p_y,t_x,t_y\\ k_1\mathrm{...}{k}_n\end{array}\right\}}{\arg \min }\underset{i=1}{\overset{n}{\&Sigma;}}\underset{j=1}{\overset{m}{\&Sigma;}}{\&lsqb;\left|Q_{i,j}{o}_i\right|-r_i\&rsqb;}^2$

Wherein, m represents the number of the point of i-th circular arc of fitting, | Q_i,jo_i| represent the point Q on i-th circular arc_i,jWith center of circle o_iIt Between distance, r_iThe radius of i-th circular arc is represented, n represents the bar number of circular arc in the conllinear circular arc group in center of circle, (p_x, p_y) represent The coordinate of the midpoint P points of line, (t between described two common intersection M points and N points_x, t_y) represent the P points to the M points or N points Vector, (k₁…k_n) represent n bar circular arcs in every circular arc proportionality coefficient；

The default nonlinear optimization algorithm is LM algorithms.

15. caliberating devices according to claim 13, it is characterised in that assume arbitrary circle in the conllinear circular arc group in arbitrary center of circle The parametrization of arc is expressed asThe center of distortion point coordinates is (x_d0, y_d0), then it is described abnormal Variable element is：

$\mathrm{\&lambda;}=\frac{1}{x_{d0}^2+y_{d0}^2+{\mathrm{Ax}}_{d0}+{\mathrm{By}}_{d0}+C}$

Wherein, A, B and C are the parameter in the parametrization expression of arbitrary circular arc.

16. caliberating devices according to claim 15, it is characterised in that the distortion peg model is：

$a=\frac{b}{1+{\mathrm{\&lambda;r}}_d^2}$

Wherein, b=(x_d, y_d) be fault image coordinate points, a=(x_u, y_u) it is image coordinate point after distortion correction, r_dFor The Euclidean distance of b and center of distortion point coordinates.