Disclosure of Invention
The invention aims to solve the problem of larger deviation of a position measurement result of the traditional PSD-based optical image motion compensation system, and can replace a fast mirror (FSM) image motion compensation mode, and provides a PSD-based two-dimensional optical fast translation image motion compensation method, a system and equipment.
The invention is realized by the following technical scheme, in one aspect, the invention provides a PSD-based two-dimensional optical rapid translational image motion compensation method, which is based on a contactless photoelectric displacement measurement device for translational image motion compensation, wherein the contactless photoelectric displacement measurement device comprises a semiconductor laser module and a one-dimensional PSD device;
The number of the one-dimensional PSD devices is 2, namely an x-axis one-dimensional PSD device and a y-axis one-dimensional PSD device, and the one-dimensional PSD devices are arranged in the lens movement direction and are respectively parallel to the x-axis and the y-axis and do not move together with the lens;
the number of the semiconductor laser modules is 2, the semiconductor laser modules are respectively matched with one-dimensional PSD devices of x and y axes to move along with the lens, and when the device is static, light beams emitted by the 2 semiconductor laser modules vertically irradiate the center position of a photosurface corresponding to the one-dimensional PSD devices;
The method specifically comprises the following steps:
Step 1, a semiconductor laser module emits light beams, the light beams vertically irradiate on a photosensitive surface of a one-dimensional PSD device of a corresponding shaft, and light spots generate photoelectric analog signals on the one-dimensional PSD device of the shaft;
Step 2, amplifying and converting the photoelectric analog signals into digital signals;
Step 3, according to the converted digital signals, position calculation, filtering and fitting correction are carried out, and final light spot position information is obtained;
Step 4, integrating the residual value of the gyroscope of the outer frame, converting the angular velocity difference into an angle difference, converting the angle difference into an radian difference, and multiplying the radian difference by the focal length value of the imaging system to obtain the linear displacement to be compensated ;
Step 5, the linear displacement of a certain shaftAs a given value, the axis fits the post-spot position informationAnd as a feedback value, comparing the feedback value with a given value in real time, controlling the lens to do two-dimensional high-speed linear translational motion along the X, Y axis according to the positive and negative of the difference value to compensate the image shift quantity, and stopping the lens after the image shift quantity compensation is completed when the difference value is zero.
Further, in step 3, the position calculation specifically includes:
for the spot position information of a one-dimensional PSD device of a certain axis, the spot offset position relative to the center position of the PSD device of the axis is obtained, and the calculation formula is as follows:
Wherein, AndIs the AD code value corresponding to the photoelectric current signal conversion of the electrodes at the two ends of the axis one-dimensional PSD, S is half of the effective length of the effective photosurface of the axis PSD device,The calculated offset relative to the center position of the PSD is the spot position information.
Further, in step 3, filtering processing is performed on the spot position information, where the filtering specifically includes:
iterative removal with IIR low pass filter The iteration process of the high-frequency noise in the light spot position information is as follows:
(1)
(2)
(3)
Wherein, Is a filter coefficient of 0<<1,Is the filtered position signal and,The old spot position value information used for iterative calculation,And calculating new spot position value information of the iteration, wherein the iteration process is repeatedly executed in sequence of formulas (1), (2) and (3).
Further, in step 3, the fitting correction is a filtered position signalThe method is characterized in that the method is generated after calibration and correction of N positions by a high-precision laser position sensor and specifically comprises the following steps:
fixing a metal cube with a receiving surface perpendicular to the direction of X, Y axis on the lens, the cube moving with the lens;
When the voice coil motor of a certain shaft drives the lens to drive the cube to move, the actual displacement of the cube is measured by the laser position sensor arranged in parallel with the shaft to correct and fit the PSD-calculated light spot position information, the whole travel range of the certain shaft is divided into N equal parts, the step length of each equal part is the total travel divided by the distance of N, and each travel sequentially runs according to the light spot position value 、、...Wherein by means ofThe position is a measurement starting point zero position, and high-precision laser position sensor values of N corresponding positions are collected simultaneously、、...After two groups of numbers are obtained, the spot position information finally generated after fitting is obtained.
Further, in step 3, the formula of the fitting correction is:
Wherein, The position information of the facula finally generated after fitting is used as position feedback of a control system,Is the filtered spot location information, A, B and C are binomial fit parameters.
Further, in step 4, the linear displacement of a certain axisThe calculation formula is as follows:
Wherein, For a given value of the angular velocity of an inertial space for a certain axis, in units of deg./s,The actual feedback angular velocity value for the gyro of this axis, in units of deg./s,In order to control the control period of the system,For the focal length of the camera, 57.3 is the angular radian conversion factor.
Further, the length of the photosurface of the one-dimensional PSD device is not smaller than the effective stroke of the lens on the corresponding axis.
In a second aspect, the invention provides a two-dimensional optical rapid translational image motion compensation system based on PSD, which comprises a non-contact photoelectric displacement measuring device, a PSD sensor signal processing board, a drive control board and a push-pull voice coil motor;
the non-contact photoelectric displacement measuring device comprises a semiconductor laser module and a one-dimensional PSD device;
The number of the one-dimensional PSD devices is 2, and the one-dimensional PSD devices are divided into an x-axis and a y-axis, are arranged in the direction of lens movement and are respectively parallel to the x-axis and the y-axis, and do not move together with the lens;
the number of the semiconductor laser modules is2, the semiconductor laser modules are respectively matched with one-dimensional PSD devices of x and y axes to move along with the lens, the setting positions are that when the device is at rest, light beams emitted by the 2 semiconductor laser modules vertically irradiate the center positions of the photosurfaces of the corresponding one-dimensional PSD devices, light spots generate photoelectric analog signals on PSDs, and the information of the center positions of the light spots relative to the PSD photosurfaces can be calculated by utilizing the signals;
The number of PSD sensor signal processing boards is 2, and the PSD sensor signal processing boards are divided into an x axis and a y axis and are used for amplifying optical position analog signals of the x axis and the y axis, converting the optical position analog signals into digital signals and transmitting the digital signals to a drive control board;
the driving control board is used for carrying out spot position calculation, filtering and fitting correction according to the digital signals after the conversion of the x axis and the y axis, obtaining the final spot position signals of the x axis and the y axis, and driving the push-pull voice coil motor to drive the lens to carry out two-dimensional linear motion;
the number of the push-pull voice coil motors is 4, wherein 2 x-axis and 2 y-axis are arranged at two ends of the lens by taking the lens as a center and are respectively parallel to the moving directions of the x-axis and the y-axis of the lens.
In a third aspect, the invention provides a computer device comprising a memory and a processor, the memory having stored therein a computer program which when executed by the processor performs the steps of a two-dimensional optical fast translational image motion compensation method based on PSD as described above.
In a fourth aspect, the present invention provides a computer readable storage medium having stored therein a plurality of computer instructions for causing a computer to perform a two-dimensional optical fast translational image motion compensation method based on PSD as described above.
The invention has the beneficial effects that:
The PSD full scale Position Sensitive detector, namely a position sensitive detector, is a photoelectric device which is used in combination with a light source and is used for detecting the accurate position of a light spot. When a long-focus optical camera images in an onboard or shipborne environment, imaging blurring is caused by movement of the image on a focal plane due to the influences of shaking of a carrier, disturbance of wind, vibration of the carrier and the like when the camera is exposed, and the imaging blurring is particularly obvious under the long-focus or long-exposure time. The introduction of the quick reflector requires that the optical lens leads out parallel light paths during design, increases the difficulty of optical design, complicates the light paths, and simultaneously has large size and high cost of the non-contact eddy current position sensor and the signal processing plate of the quick reflector.
Compared with the traditional rapid reflector design method, the invention has the advantages that 1) the translational motion of the high-speed linear lens replaces the rotational motion of the reflector, a parallel light path is not needed to be specially designed, the optical design is simple, and 2) the PSD position sensor replaces an eddy current sensor, and the invention has low cost, small size and compact integral structure.
The invention is suitable for a secondary image motion compensation system of a long-focus photoelectric imaging device with strict structural size requirement and simple optical design. Due to the disturbance effects of carrier vibration, atmospheric disturbance, attitude change and the like, the gyro compensation mechanism cannot completely compensate image movement caused by disturbance, and a secondary image movement compensation system is required to be introduced for further image stabilization.
Detailed Description
Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals refer to the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the drawings are exemplary and intended to illustrate the present invention and should not be construed as limiting the invention.
In the first embodiment, a two-dimensional optical rapid translational image motion compensation method based on PSD is based on a contactless photoelectric displacement measurement device for translational image motion compensation, wherein the contactless photoelectric displacement measurement device comprises a semiconductor laser module 1 and a one-dimensional PSD device 2;
the number of the one-dimensional PSD devices 2 is 2, namely an x-axis and a y-axis one-dimensional PSD devices 2 respectively, and the one-dimensional PSD devices 2 are arranged in the lens movement direction and are respectively parallel to the x-axis and the y-axis and do not move along with the lens;
the number of the semiconductor laser modules 1 is 2, the semiconductor laser modules are respectively matched with the one-dimensional PSD devices 2 of the x-axis and the y-axis and move together with the lens, the setting positions are that when the device is at rest, light beams emitted by the 2 semiconductor laser modules 1 vertically irradiate the center position of a photosurface corresponding to the one-dimensional PSD device 2, light spots generate photoelectric analog signals on PSDs, and the position information of the light spots relative to the center of the photosurface of the PSD can be calculated by utilizing the signals;
The method specifically comprises the following steps:
Step 1, a semiconductor laser module 1 emits light beams, the light beams vertically irradiate on a photosurface of a one-dimensional PSD device 2 of a corresponding shaft, and light spots generate photoelectric analog signals on the PSD device of the shaft;
step 2, amplifying and converting the photoelectric analog signals into digital signals in a PSD sensor signal processing board 3;
step 3, in the drive control board 4, according to the converted digital signals, position calculation, filtering and fitting correction are carried out, and final x-axis and y-axis light spot position information is obtained;
And 4, calculating an image displacement value to be compensated by using a known imaging system image displacement compensation formula, namely, calculating the linear displacement value of the lens to be moved at the X, Y axis. Firstly, integrating the residual value of the gyro of the outer frame to convert the angular velocity difference into an angle difference, the calculating method is to subtract the gyro feedback velocity value from the given value of the inertial space velocity of the outer frame, then multiplying the angle difference by the control time to obtain the angle difference, then converting the calculated angle difference into the radian difference by utilizing the relation formula that the arc length of a circle is equal to the radian multiplied by the diameter, then multiplying the radian difference by the focal length value of the imaging system to become the linear displacement to be compensated, the calculating method of the x-axis and the y-axis is the same, and the linear displacement of the image displacement of a certain axis is the same The calculation formula is as follows:
Wherein the method comprises the steps of For a given value of the angular velocity of an inertial space for a certain axis, in units of deg./s,The actual feedback angular velocity value for the gyro of this axis, in units of deg./s,In order to control the control period of the system,For the focal length of the camera, 57.3 is an angle radian conversion coefficient, and the calculated result converts the angle rotation quantity required for image shift compensation into a linear displacement quantity。
Step 5, the image displacement and linear displacement of the outer frame compensation of one axisAs a given value, the axis is fitted toThe light spot position information is used as a feedback value, the feedback value is compared with a given value in real time, the output control of the driver is adjusted to drive the voice coil motor 5 to operate according to the comparison result, the lens is driven to do two-dimensional high-speed linear translational motion along the X, Y axis to compensate image shift, namely, the feedback value is compared with the given value in real time, the lens is controlled to do two-dimensional high-speed linear translational motion along the X, Y axis according to the positive and negative of the difference value and the size to compensate the image shift, and when the difference value is zero, the lens stops moving after the image shift compensation.
The embodiment provides a two-dimensional optical fast translational image motion compensation method based on PSD, which is used for reducing image motion generated during imaging of a long-focal-length camera, and replaces two-dimensional rotation of an FSM (fast mirror) originally used by utilizing two-dimensional linear motion of a lens. Compared with the traditional design method of the quick reflector, the design method of the quick reflector adopts the high-speed linear translational motion of the lens to replace the rotational motion of the reflector, does not need to specially design a parallel light path, has simple optical design, and has low cost, small size and compact overall structure because the PSD position sensor replaces the eddy current sensor.
In a second embodiment, the present embodiment further defines a two-dimensional optical fast translational image motion compensation method based on PSD, where step 3 is further defined, and specifically includes:
in step 3, the position calculation specifically includes:
for the spot position information of the one-dimensional PSD device 2 of a certain axis, the spot offset position relative to the center position of the PSD of the axis is obtained, and the calculation formula is as follows:
Wherein, AndIs the AD code value corresponding to the photoelectric current signal conversion of the electrodes at the two ends of the axis one-dimensional PSD, S is half of the effective length of the effective photosurface of the axis PSD device,The calculated offset relative to the central position of the PSD is spot position information, and the spot position information needs to be filtered, fitted and corrected to further improve the precision.
In this embodiment, the spot position information is calculated by calculating the offset of the spot relative to the center point of the one-dimensional PSD of the axis according to the voltage code value output by the electrodes at both ends of the one-dimensional PSD.
In a third embodiment, the present embodiment further defines a two-dimensional optical fast translational image motion compensation method based on PSD, where step 3 is further defined, and specifically includes:
in step 3, the spot position information needs to be filtered, where the filtering specifically includes:
iterative removal with IIR low pass filter The iteration process of the high-frequency noise in the light spot position information is as follows:
(1)
(2)
(3)
Wherein, Is a filter coefficient of 0<<1,Is the filtered position signal and,The old spot position value information used for iterative calculation,And calculating new spot position value information of the iteration, wherein the iteration process is repeatedly executed in sequence of formulas (1), (2) and (3).
In the embodiment, the IIR low-pass filter is adopted to filter out high-frequency noise of the system, so that the calculated amount occupies little MCU resource and is easy to realize in a digital control system.
In a fourth embodiment, the present embodiment further defines a two-dimensional optical fast translational image motion compensation method based on PSD, where step 3 is further defined, and specifically includes:
in step 3, the fitting correction is a filtered position signal The method is characterized in that the method is generated after calibration and correction of N positions by a high-precision laser position sensor and specifically comprises the following steps:
fixing a metal cube with a receiving surface perpendicular to the direction of X, Y axis on the lens, the cube moving with the lens;
When the voice coil motor of a certain shaft drives the lens to drive the cube to move, the actual displacement of the cube is measured by the laser position sensor arranged in parallel with the shaft to correct and fit the PSD-calculated light spot position information, the whole travel range of the certain shaft is divided into N equal parts, the step length of each equal part is the total travel divided by the distance of N, and each travel sequentially runs according to the light spot position value 、、...Wherein by means ofThe position is a measurement starting point zero position, and high-precision laser position sensor values of N corresponding positions are collected simultaneously、、...And after the two groups of numbers are obtained, acquiring spot position information finally generated after fitting.
The fitting correction of the embodiment has the effect of improving the position accuracy of the speckles, and further improving the accuracy of image motion compensation.
In a fifth embodiment, the present embodiment further defines a two-dimensional optical fast translational image motion compensation method based on PSD, where step 3 is further defined, and specifically includes:
In step 3, fitting correction is required to be performed on the filtered spot position information, and the fitting correction formula specifically comprises:
Wherein, The position information of the facula finally generated after fitting is used as position feedback of a control system,Is the filtered spot location information, A, B and C are binomial fit parameters.
In this embodiment, the function of fitting and correcting the spot position information is to improve the spot position accuracy, and further improve the accuracy of image motion compensation.
In a sixth embodiment, the present embodiment further defines a two-dimensional optical fast translational image motion compensation method based on PSD, where step 4 is further defined, and specifically includes:
in step 4, the linear displacement of one axis The calculation formula is as follows:
Wherein, For a given value of the angular velocity of an inertial space for a certain axis, in units of deg./s,The actual feedback angular velocity value for the gyro of this axis, in units of deg./s,In order to control the control period of the system,For the focal length of the camera, 57.3 is the angular radian conversion factor.
The present embodiment provides a specific calculation formula of the linear displacement amount of a certain axis.
In a seventh embodiment, the present embodiment further defines a two-dimensional optical fast translational image motion compensation method based on PSD, where the one-dimensional PSD device is further defined, and specifically includes:
the length of the photosurface of the one-dimensional PSD device 2 is not smaller than the effective stroke of the lens on the corresponding axis.
The function of this embodiment is to ensure that the light beam emitted from the semiconductor laser module 1 can form a light spot on the photosensitive surface of the one-dimensional PSD device 2 in the entire effective stroke of the lens.
Embodiment eight, this embodiment is an embodiment 1 of a two-dimensional optical fast translational image motion compensation method based on PSD, which specifically includes:
As shown in fig. 1, the method specifically includes the following steps:
Step 1, a semiconductor laser module 1 emits light beams, the light beams vertically irradiate on a photosensitive surface of a PSD device 2 of a corresponding shaft, and photoelectric analog signals are generated on the PSD device of the shaft by light spots;
step 2, amplifying and converting the photoelectric analog signals into digital signals in a PSD sensor signal processing board 3;
step 3, in the drive control board 4, according to the converted digital signals, position calculation, filtering and fitting correction are carried out, and final light spot position information is obtained;
And 4, calculating an image displacement value to be compensated by using a known imaging system image displacement compensation formula, namely, calculating the linear displacement value of the lens to be moved at the X, Y axis. Firstly, integral operation is carried out on the residual value of the gyroscope of the outer frame, and the angular velocity difference value is converted into an angle difference value. The method comprises subtracting the feedback value of gyro speed from the given value of inertial space speed of the outer frame, multiplying by control time to obtain an angle difference value, converting the calculated angle difference value into an arc difference value by utilizing a relation formula of arc length equal to radian multiplied by diameter of a circle, multiplying by the focal length value of the imaging system to become the linear displacement to be compensated, and calculating the displacement of a certain axis by the same method as that of the x-axis and the y-axis The calculation formula is as follows:
Wherein the method comprises the steps of For a given value of the angular velocity of an inertial space for a certain axis, in units of deg./s,The actual feedback angular velocity value for the gyro of this axis, in units of deg./s,In order to control the control period of the system,For the focal length of the camera, 57.3 is an angle radian conversion coefficient, and the calculated result converts the angle rotation quantity required for image shift compensation into a linear displacement quantity。
Step 5, the image shift amount of the outer frame compensation of one axisAs a given value, the axis after fittingThe light spot position information is used as a feedback value, the feedback value is compared with a given value in real time in a drive control board 4, and according to the comparison result, the output of the driver is adjusted to control the push-pull voice coil motor 5 to run so as to drive the lens to do two-dimensional high-speed linear translational motion along the X, Y axis to compensate image movement.
In step 3, the position resolving specifically includes:
for the spot position information of the one-dimensional PSD device 2 of a certain axis, the spot offset position relative to the center position of the PSD of the axis is obtained, and the calculation formula is as follows:
Wherein, AndIs the AD code value corresponding to the photoelectric current signal conversion of the electrodes at the two ends of the axis one-dimensional PSD, S is half of the effective length of the effective photosurface of the PSD device,The calculated offset relative to the central position of the PSD is spot position information, and the spot position information needs to be filtered and fitted to further improve the precision.
In step 3, the spot position information needs to be filtered, where the filtering specifically includes:
iterative removal with IIR low pass filter The iterative process is as follows:
(1)
(2)
(3)
Wherein, Is a filter coefficient of 0<<1,Is the filtered position signal and,The old spot position value information used for iterative calculation,And calculating new spot position value information of the iteration, wherein the iteration process is repeatedly executed in sequence of formulas (1), (2) and (3).
In the step 3, correction fitting is required to be performed on the filtered spot position information, and the fitting correction formula specifically includes:
Wherein, Is the spot position information after a certain axis filtering, A, B and C are binomial fitting parameters,The final generated spot position information after fitting is used as the position feedback of the linear motion of the shaft, and the fitting correction function is to improve the spot position precision and further improve the image motion compensation precision.
An embodiment nine, a two-dimensional optical fast translational image motion compensation system based on PSD, specifically includes:
as shown in fig. 2, the implementation system comprises a non-contact photoelectric displacement measuring device, a PSD sensor signal processing board 3, a drive control board 4 and a push-pull voice coil motor 5;
The non-contact photoelectric displacement measuring device comprises a semiconductor laser module 1 and a PSD device 2;
the number of PSD devices 2 is 2, and the PSD devices are divided into an x-axis and a y-axis one-dimensional PSD devices 2, and are arranged in the direction of lens movement and are respectively parallel to the x-axis and the y-axis and do not move together with the lens;
The number of the semiconductor laser modules 1 is 2, the semiconductor laser modules are respectively matched with the one-dimensional PSD devices 2 of the x-axis and the y-axis to move along with the lens, the setting positions are that when the device is at rest, light beams emitted by the 2 semiconductor laser modules 1 vertically irradiate the center positions of the photosurfaces of the corresponding one-dimensional PSD devices 2, light spots generate photoelectric analog signals on PSDs, and the information of the center positions of the light spots relative to the PSD photosurfaces can be calculated by utilizing the signals;
The number of the PSD sensor signal processing boards 3 is 2, and the PSD sensor signal processing boards 3 are divided into an x axis and a y axis and are used for amplifying optical position analog signals of the x axis and the y axis, converting the optical position analog signals into digital signals and transmitting the digital signals to the drive control board 4;
The driving control board 4 is used for carrying out spot position calculation, filtering and fitting correction according to the digital signals after the conversion of the x axis and the y axis, obtaining the final spot position signals of the x axis and the y axis, and driving the push-pull voice coil motor 5 to drive the lens to carry out two-dimensional linear motion;
The number of push-pull voice coil motors 5 is 4, wherein 2 x-axis and 2 y-axis are arranged at two ends of the lens with the lens as a center and are respectively parallel to the moving directions of the x-axis and the y-axis of the lens. When the lens moves along an axis, one voice coil motor of the axis is responsible for pushing, one voice coil motor is responsible for pulling, the two motors jointly drive the lens to complete linear motion along the axis, and the push-pull driving has the advantages that stress is balanced when the lens moves linearly, and high-frequency reciprocating motion is easy to realize.
The light beam emitted by the semiconductor laser module 1 is vertically irradiated on the photosensitive surface of the one-dimensional PSD device 2 of the corresponding shaft, the formed light spot generates a photoelectric analog signal on the PSD device, the analog signal is amplified and AD converted into a digital signal through the PSD sensor signal processing board 3, the digital signal is subjected to position calculation, filtering and fitting correction in the driving control board 4 to generate light spot position information, and meanwhile the driving control board 4 drives the push-pull voice coil motor 5 to drive the lens to linearly move in two dimensions.
The lens and the semiconductor laser module 1 are installed together to form a moving part, and the PSD device 2, the PSD signal processing board 3 and the drive control board 4 are stationary parts, so that light spots can synchronously move on a PSD light sensitive surface along an axis when the lens moves along the axis.
A PSD-based two-dimensional optical rapid translational image shift compensation method utilizing the system of the embodiment comprises the following steps:
step 1, a semiconductor laser module 1 emits light beams, the light beams vertically irradiate on a photosurface of a one-dimensional PSD device 2 of a corresponding shaft, photoelectric analog signals generated by light spots on the one-dimensional PSD device 2 require that the photosurface length of the one-dimensional PSD device 2 is not less than the effective stroke of a lens of the shaft;
step 2, amplifying and converting the photoelectric analog signals into digital signals in a PSD sensor signal processing board 3;
Step 3, position calculation, filtering and fitting correction are carried out in the drive control board 4 according to the converted digital signals, and final light spot position information is obtained;
And 4, calculating an image displacement value to be compensated by using a known imaging system image displacement compensation formula, namely, calculating the linear displacement value of the lens to be moved at the X, Y axis. Firstly, integral operation is carried out on the residual value of the gyroscope of the outer frame, and the angular velocity difference is converted into an angle difference value. The method comprises subtracting gyro feedback speed value from inertial space speed set value of outer frame, integrating the obtained angular speed difference by multiplying control time to obtain angle difference, converting calculated angle difference into radian difference by using relation formula of arc length equal to radian multiplied by diameter of circle, multiplying focal length value of imaging system to obtain linear displacement to be compensated, calculating x-axis and y-axis, and calculating image displacement of one axis The calculation formula is as follows:
Wherein the method comprises the steps of For an axis, the angular velocity of the inertial space is given by the unit of DEG/s,The actual feedback angular velocity value for the gyro of this axis, in units of deg./s,In order to control the control period of the system,For the focal length of the camera, 57.3 is an angle radian conversion coefficient, and the calculated result converts the angle rotation quantity required for image shift compensation into a linear displacement quantity。
Step 5, the image shift amount of the outer frame compensation of one axisAs a given value, after fittingThe light spot position information is used as a feedback value, the feedback value is compared with a given value in real time, and according to the comparison result, the output of the driver is adjusted to control the push-pull voice coil motor 5 to run so as to drive the lens to do two-dimensional high-speed linear translational motion along the X, Y axis to compensate image movement.
In step 3, the light spot position calculation is completed in the driving control board 4, which specifically includes:
for the spot position information of the one-dimensional PSD device 2 of a certain axis, the spot offset position relative to the center position of the PSD of the axis is obtained, and the calculation formula is as follows:
Wherein, AndIs the AD code value corresponding to the photoelectric current signal conversion of the electrodes at the two ends of the axis one-dimensional PSD, S is half of the effective length of the effective photosurface of the PSD device,The calculated offset relative to the PSD center position is the spot position information, and the spot position information at the moment needs to be filtered and fitted to be calculated to further improve the precision.
In step 3, the spot position information needs to be filtered, where the filtering specifically includes:
iterative removal with IIR low pass filter The iterative process is as follows:
(1)
(2)
(3)
Wherein, Is a filter coefficient of 0<<1,Is the filtered position signal and,The old spot position value information used for iterative calculation,And calculating new spot position value information of the iteration, wherein the iteration process is repeatedly executed in sequence of formulas (1), (2) and (3).
In step 3, the filtered spot position information needs to be corrected and fitted, and the fitting correction is completed in the drive control board 4 and used as final spot position information.
The modified fit is filteredThe position signals are generated after N position calibration and correction by the high-precision laser position sensor. The method is that a metal cube with a receiving surface perpendicular to the X, Y axis direction is fixed on the lens, the cube moves together with the lens, as shown in figure 3, when the voice coil motor of a certain axis drives the lens to drive the cube to move, the actual displacement of the cube is measured by a laser position sensor placed parallel to the axis to correct and fit the PSD-resolved spot position information, the whole range of a certain axis is divided into N equal parts, the step length of each equal part is the total travel divided by the distance of N, namely each travel sequentially moves according to the spot position value、、...Wherein by means ofThe position is a measurement starting point zero position, and high-precision laser position sensor values of N corresponding positions are collected simultaneously、、...After two groups of numbers are obtained, a quadratic polynomial generated by using mathematical tools such as Origin or MATLAB is expressed as follows:
Wherein, Is the spot position information after a certain axis filtering, A, B and C are binomial fitting parameters,The final generated spot position information after fitting is used as the axial position feedback, and the fitting correction function is to improve the spot position precision and further improve the image motion compensation precision.
Due to the influences of factors such as installation errors and nonlinearity of the PSD device, the calculated light spot position and the actual value deviate, and new light spot position information is generated by fitting and correcting after the calibration of a high-precision position sensor and is used as a feedback value. The fitting correction calculation is completed in the drive control board 4, the position accuracy of the position sensor for fitting calibration is higher than that of PSD, and one embodiment uses an ohm-dragon ZW-S20 laser displacement sensor for calibration correction.
The operation amplifier needed by weak signal detection of the light spot current in the PSD sensor signal processing board 3 is required to have low noise, high bandwidth and low bias current, the AD device is selected to have high resolution, good linearity and can synchronously sample the voltage converted at two ends of the one-dimensional PSD, wherein the AD device adopts AD7606 with 16-bit resolution synchronously sampled by the AD device, and the low bias current and high bandwidth instrument operation amplifier of the LT1464A of the ADI company is selected by the amplifier of one embodiment.
In one embodiment, the driving control board 4MCU is adopted to support hardware floating point operation, and the megaworkable ARM Cortex-M4 microcontroller GD32F450 with the main frequency of up to 200 MHz.
Testing the image motion compensation effect of the tele camera under the disturbance condition of 2 degrees@1 Hz of a swinging table, and when the outer frame gyro is subjected to rough compensation, inputting the residual image motion value as a two-dimensional optical rapid translation image motion compensation input, wherein the root mean square value of the image motion at the moment of exposure after compensation is controlled within 0.6 image elements, so that the requirement that the image motion at the moment of imaging exposure is not more than 1 image element index is met.