CN101513342A - Full-view pupil analysis measurement method - Google Patents

Abstract

The present invention is a full-view pupil analysis and measurement method, comprising the following steps: step 300, first collect the image data of the eye region in real time through a video capture card, and obtain an image of the eye region; step 301, through the mean value method or the maximum value method Or the formula method carries out gray scale to the image that 300 steps collect; Step 302, carries out binarization processing to the image of step 301 gray scale by the histogram method, when background and target difference are big, the histogram appears bimodal, this The lowest valley point during the time is used as the threshold point; step 303, eliminate the reflection of the image pupil area of the binarized image; step 304 determines the center of the pupil, calculates its diameter, extracts the center of the pupil, and determines the diameter of the pupil to realize the center of the pupil tracking measurement. This method is a full-field pupillometric analysis method with good real-time performance and convenient use.

Description

Full field pupil analysis measurement method

technical field

The invention belongs to image detection and processing technology, in particular to a full-view pupil analysis and measurement method.

Background technique

At present, with the continuous advancement of science and technology and medical level, studies have shown that objective and quantitative pupil detection is more and more applied in various directions.

In medical research, excimer laser refractive surgery (LASIK), neuro-ophthalmology (such as retrobulbar optic neuritis), diabetes, neurobiology and autonomic nervous system function research, central nervous system diseases, brain function, neuropsychiatric diseases ( Such as monitoring of increased intracranial pressure, observation of cerebral apoplexy, auxiliary judgment of brain death), intensive care and emergency medicine, anesthesia (such as anesthesia depth monitoring), clinical pharmacology, eye movement detection, video nystagmus detection/infrared nystagmus detection (using Vertigo, vestibular disease inspection), etc.; drug abuse research and testing, can be used in scientific research and testing fields such as the definition of drug abuse, the determination of the degree of physical and psychological addiction (craving), and the evaluation of drug rehabilitation effects. Appraisal institutions and related scientific research institutions can greatly reduce the cost of verification, and greatly save the initial screening time of drugs and other drugs; the judgment of fatigue and drinking, the use of pupil detectors in industry, transportation, transportation and other departments can effectively evaluate workers or A driver's level of fatigue (drowsy driving) or the presence or absence of drunk driving can help reduce the occurrence of related accidents and drunk crashes, and can improve workplace safety and increase work efficiency.

The existing simple pupil measuring boards in China cannot achieve real-time measurement, and the binocular pupil detector is not suitable for carrying. Other hand-held pupil measuring devices are easy to carry, but cannot obtain real-time measurement results.

Contents of the invention

The object of the present invention is to provide a full-field pupil analysis measurement method with good real-time performance and convenient use.

The purpose of the present invention is achieved by the following technical solutions: a full-field pupil analysis measurement method, comprising the steps:

Starting from step 300, firstly, the image data of the eye area is collected in real time through a video capture card to obtain an image of the eye area;

Step 301 grayscales the image collected in step 300 by the mean value method or the maximum value method or the formula method;

Step 302, binarize the grayscaled image in step 301 by the histogram method, when the difference between the background and the target is large, the histogram has double peaks, and the lowest valley point at this time is used as the threshold point;

Step 303, eliminating the reflection in the pupil region of the image undergoing binarization processing;

Step 304 determines the center of the pupil, calculates its diameter, extracts the center of the pupil, and determines the diameter of the pupil, so as to realize the tracking measurement of the pupil center.

The method for determining the center of the circle can be determined by the intersection of the perpendicular bisectors of the three chords.

The method for determining the center of the circle can be determined by the midpoints of two chords or by three points on the circle, and further determined by the following formula: Let (a, b) be the center of the circle, R be the radius, (x1, y1), ( x2, y2), (x3, y3) are three points on the circle. According to the following formula, the values of a, b, and R can be solved finally, and then the center of the circle can be determined and the area can be calculated.

(x1-a) ² +(y1-b) ² ＝R ²

(x2-a) ² +(y2-b) ² ＝R ²

(x3-a) ² +(y3-b) ² =R ² .

In the step 301, the gray scale of the same color image is processed by using the formula: Y=0.3R+0.59G+0.11B.

The device adopted in the method at least includes an image detection unit and an image processing unit, and the image detection unit and the image processing unit are electrically connected.

The electrical connection may be a wired electrical connection.

Said electrical connection may be a radio connection.

The image detection unit is electrically connected with a wireless transmitting device, and the wireless transmitting device is a video wireless transmitting device.

The image processing unit is electrically connected with a wireless receiving device, and the wireless receiving device is a video receiving card.

The image processing unit is a notebook computer.

The image detection unit is a CMOS camera or a CCD camera.

In the present invention, three points are used to determine the center of the circle and the method of scanning pixels to obtain the diameter is adopted.

Working process of the present invention and advantage are:

Using wired video or wireless transmitting device and video receiving card for video information transmission, the hand-held pupil measuring instrument composed of camera can be transmitted to the notebook computer in real time with the acquisition frequency of 30 frames per second, and the image processing software of the notebook computer Quickly analyze the pupil image and get the analysis result quickly.

Since the existing notebook computer is easy to carry, and the notebook computer has a fast image processing speed, it can realize the tasks of real-time measurement and image analysis.

Description of drawings

The present invention will be further described below in conjunction with the accompanying drawings of the embodiments.

Fig. 1 is a structural representation of the present invention's implementation 1;

Fig. 2 is the structure schematic diagram of implementing 2 of the present invention;

Fig. 3 is pupil image extraction and rapid analysis and flowchart;

Fig. 4 is to obtain eye region image;

Fig. 5 is a schematic diagram of an image after binarization;

Fig. 6 is a schematic diagram of a pupil reflection image after binarization;

Fig. 7 is a principle process diagram of expansion in image processing;

Fig. 8 is to convert the region of the pupil into a circular domain (binary image);

FIG. 9 is to extract the boundary of the circle domain obtained after binarization (extract the contour of the circle domain).

In the figure: 1. Image processing unit; 2. Image detection unit; 3. Radio; 4. Wireless receiving device; 5. Wireless transmitting device; 6. Handle; 7. Switch; 8. Camera; 9. Lighting lamp; 10. Driving circuit; 11. Connecting wire; 12. Video interface terminal.

Detailed ways

As shown in Figure 1, this is an example method that utilizes an existing notebook computer with a camera to improve, and is different from an existing notebook computer with a camera, the original camera 8 of the notebook computer (image processing unit 1) is connected The line 11 is electrically connected with the notebook computer; the notebook computer includes a video interface terminal 12 . In this way, the image detection unit 2 should include a camera 8, a drive circuit 10, a housing for fixing the camera 8 and the drive circuit 10, a handle 6 is arranged at the lower end of the housing, and a switch 7 is arranged on the handle 6. The image detection unit 2 actually constitutes a hand-held pupillometer.

The power supply of the image detection unit 2 is directly taken from the notebook computer power supply; the original camera 8 drive circuit 10 of the notebook computer is moved from the image detection unit 2, so that the image detection unit 2 directly outputs the video signal to the notebook computer, and then the notebook computer video signal processing.

Since this implementation mode does not require creative labor, it is particularly convenient to implement. In order to realize the needs of the present invention for pupil measurement and processing, there is image processing software in the notebook computer, through the gray scale processing of the pupil image extracted by the camera 8, edge extraction, and scanning pixels to analyze the single frame image of the video, it can be eliminated or compensated. The uncertainty of the pupil diameter due to the error caused by blinking ensures the reliability and repeatability of the measurement results.

In order not to affect the inspection in a dark environment, the side of the camera 8 is equipped with a light source, which is a lighting lamp 9, and the light source can be an external retractable light source that can be controlled, including ensuring that no light leaks to another pupil when irradiating a single pupil, which will affect the measurement results. ;Ensure that it is not affected by indoor and outdoor lighting conditions when checking in a dark environment, and ensure the consistency of testing in various environments. The camera 8 can adjust and calibrate the object distance. When the eyes of different patients are different due to the degree of convexity and concaveness, the object distance of the camera can be adjusted separately according to the needs, so as to ensure that the best individual image quality is always obtained, and the measurement results are not affected by the object distance. impact of changes.

When using it, first align the hand-held pupil measuring instrument with the eye of the person under test (both left and right eyes are acceptable), turn on the switch, adjust the distance through the focus button, and get a clear video image of the eye under test on the screen, press the start button, The recording starts, and when the recording ends, press the stop button.

Fig. 2 is a kind of wireless implementation scheme, in this scheme, by connecting a wireless transmitting device 5 at the video output end of image detection unit 2 (hand-held pupil measuring instrument), wireless transmitting device 5 is a video wireless transmitting device , continuously sending out wireless video signals through radio 3. On the USB interface of the notebook computer (image processing unit 1), a wireless receiving device 6 is electrically connected, and the wireless receiving device 6 is a video receiving card. The long-distance interaction between the image detection unit 2 (handheld pupil measuring instrument) and the notebook computer (image processing unit 1) can be realized.

Both Embodiment 1 and Embodiment 2 utilize the existing technology to make structural changes, so that the full-field pupil analysis becomes convenient and accurate. And (handheld pupil measuring instrument) and notebook computer (image processing unit 1) are the key to realize this technology through wired or wireless connection.

Fig. 3 is to utilize Fig. 1 or Fig. 2 to carry out pupil image extraction and fast analysis and flowchart:

Step 300 is the start of the pupil extraction process. Firstly, the image data of the eye area is collected in real time through a video capture card to obtain an image of the eye area, as shown in FIG. 4 .

Step 301 is the grayscale processing of the image. The image collected in step 300 is grayscaled by the mean value method or the maximum value method or the formula method. The above three methods grayscale the same color image. Observation and comparison found that the effect of the image processed by the formula method is ideal, the color of the image is relatively more uniform, and the visual effect is better. Therefore, this software adopts the third method for grayscale conversion, that is, using the formula: Y=0.3R+ 0.59G+0.11B.

The grayscaled image in step 301 is binarized by the histogram method. The histogram method is that when the difference between the background and the target is large and the area is equal, the histogram appears double peaks. At this time, the lowest valley point is T, so when the binarization is performed, the lowest valley point T is used as the threshold point, and step 302 is performed. Binarization processing of the inner region, as shown in Figure 5.

Step 303 Elimination of reflections in the pupil area. When the eye image is collected in real time, a reflection area will be generated on the pupil. After binarization, the reflection area can be clearly seen. As shown in the white area in Figure 5, these reflection areas will affect The next step is to locate the center of the pupil, so the reflective area must be eliminated.

为B的关于原点对称的集合，

表示平移z. $A\⊕B=\left\{z\right|[{\left(\hat{B}\right)}_Z\∩A]\⊆A\}.$ B膨胀A实际上就是

的位移与A至少有一个非零元素相交时B的原点为位置的集合。As shown in Figure 7, the elimination of the reflective area is based on the principle of expansion in image processing. First a simple definition of inflation: Let A be the set of image regions and B be the structural element.

is a set of B that is symmetric about the origin,

express translation z. $A\⊕B=\left\{z\right|[{\left(\hat{B}\right)}_Z\∩A]\⊆A\}.$ B inflate A is actually

The set of positions where the origin of B intersects with at least one nonzero element of A.

According to Figure 7, it can be seen that if you want to automatically fill the blank inside the image with black, the key must be to distinguish the white of the background from the white of the object (the inner area of the pupil). The solution to this problem is as follows: First, scan line by line to find out the pixels that are black on the left and white on the right, so that the white area of the background on the left can be excluded, and the remaining area is shown in Figure 6. Red, blue mark. Secondly, judge whether there are black pixels on the right side of the row where each pixel in the red area is located. If there are black pixels, it means that this point is inside, otherwise it can be judged that this point is the background of the blue area. White. Finally, the determined internal area is automatically filled. The result is shown in Figure 8 below.

Step 304 determines the center of the pupil and calculates its diameter. The ultimate purpose of the embodiment of the present invention is to extract the center of the pupil and determine the diameter of the pupil (the pupil can be approximated as a circle temporarily), so as to realize the tracking measurement of the pupil center. According to geometric methods, there are many ways to determine the center of a circle. It can be determined by the intersection of the perpendicular bisectors of the three chords, the midpoint of the two chords, or the three points on the circle, and further use formulas to determine. In the present invention, three points are used to determine the center of the circle and the method of scanning pixels to obtain the diameter is adopted.

Calculate the diameter from the center of a circle:

Let (a, b) be the center of the circle, R be the radius, and (x1, y1), (x2, y2), (x3, y3) be three points on the circle. According to the following formula, the values of a, b, and R can be solved finally, and then the center of the circle can be determined and the area can be calculated.

(x1-a) ² +(y1-b) ² ＝R ²

(x2-a) ² +(y2-b) ² ＝R ²

(x3-a) ² +(y3-b) ² ＝R ²

According to the above formula, how to get the three points on the circle is the key. After the original image has been processed through the above process, the area of the pupil has been converted into a circular domain (binary image) as shown in FIG. 8 . So how to get the point on the boundary of the circle domain is the key. Use the following method to get three points on the circle. First, the boundary extraction (extraction of the contour of the circle) of the circle domain obtained after binarization is performed as shown in Figure 9. After the circle is obtained, three different points are taken on the circle, and then the center of the circle is calculated according to the above formula. Radius, and then determine the center of the circle, the diameter can be calculated.

Contour extraction determines 3 points on the circle:

The main implementation principle of contour extraction is to scan the image line by line to find out the point of black pixels, and then further judge that the top, bottom, left and right of this point are all black pixels, if so, attach this black pixel to White, otherwise do not change this pixel. Finally, the center point of the pupil can be determined by 3 points on the edge, and the diameter of the pupil can be calculated by extracting the coordinates of the points on the edge.

Scan pixels to get diameter:

On the edge-extracted image, scan the pixels from the bottom of the image line by line, and record the ordinate and abscissa of the first black pixel scanned, and continue to scan vertically upwards along the abscissa recorded at this point until it is scanned For the second black pixel, record the ordinate of this point, the difference between the ordinates of the two points is the pupil diameter, and the unit is pixel, so the pupil diameter can be obtained.

After each value is measured, it is judged by the software. If the difference between the measured value and the first 5 and the last 5 measured values exceeds 20% of the average value of 10, then it is judged that the value is a blink wave, and it is excluded. The measurement results are compensated.

Claims (10)

1, full-view pupil Measurement and analysis method comprises the steps:

From step 300, at first the view data of ocular is gathered in real time by video frequency collection card, obtain the ocular image;

Step 301 is by averaging method or be worth method most or equation carries out gray processing to 300 images that collect of step;

Step 302 is carried out binary conversion treatment by histogram method to the image of step 301 gray processing, and when background and target difference are big, rectangular histogram occurs bimodal, and minimum valley point at this moment is as threshold point;

Step 303 is to carrying out reflective elimination of image pupil region of binary conversion treatment;

Step 304 is determined pupil center, calculates its diameter, extracts the center of circle of pupil, and the diameter of definite pupil, realizes the tracking measurement of pupil center.

2, full-view pupil Measurement and analysis method according to claim 1 is characterized in that: the method in described definite center of circle can be passed through the intersection point of the perpendicular bisector of three strings and determine.

3, full-view pupil Measurement and analysis method according to claim 1, it is characterized in that: the method in described definite center of circle can the mid point by two strings be determined or by 3 points on the circle, further utilize following formula to determine: to establish that (a b) is the center of circle, and R is a radius, (x1, y1), (x2, y2), (x3 y3) is 3 points on the circle; According to following formula, finally can solve a, b, the value of R, and then can determine the center of circle, calculate area:

(x1-a) ²+(y1-b) ²＝R ²

(x2-a) ²+(y2-b) ²＝R ²

(x3-a) ²+(y3-b) ²＝R ²。

4, full-view pupil Measurement and analysis method according to claim 1 is characterized in that: it is to utilize formula that described step 301 pair same width of cloth coloured image carries out gray processing: Y=0.3R+0.59G+0.11B handles.

5, full-view pupil measurement analysis device which comprises at least an image detecting element (2), a graphics processing unit (1), it is characterized in that: pass through between image detecting element (2) and the graphics processing unit (1) to be electrically connected.

6, full-view pupil measurement analysis device according to claim 5 is characterized in that: described electrical connection can be wired electrical connection or dedicated radio link.

7, full-view pupil measurement analysis device according to claim 5 is characterized in that: described image detecting element (2) is electrically connected with wireless launcher (5), and wireless launcher (5) is a video wireless launcher.

8, full-view pupil measurement analysis device according to claim 5 is characterized in that: described graphics processing unit (1) is electrically connected with radio receiver (4), and radio receiver (4) is the video reception card.

9, full-view pupil measurement analysis device according to claim 5 is characterized in that: described graphics processing unit (2) is a notebook computer.

10, full-view pupil measurement analysis device according to claim 5 is characterized in that: described image detecting element (2) is cmos camera or ccd video camera.

Images