A method, system and medium for removing abnormal image pixels

Abstract

The present invention discloses a method, system and medium for removing abnormal image pixels. The removal method includes: confirming the position of abnormal pixel points corresponding to vertical stripes in a photosensitive pixel row; dividing multiple exposure intervals, and corresponding a number of exposure times to each exposure interval, so as to obtain the grayscale value corresponding to each exposure time; fitting each exposure time and its corresponding grayscale value in all exposure intervals, so as to obtain the compensation gain coefficient corresponding to all exposure intervals; accumulating and calculating the interval compensation values ​​corresponding to all exposure intervals covered between the real-time exposure time and the maximum exposure time or the minimum exposure time, so as to obtain the grayscale compensation value corresponding to the abnormal pixel point. The present invention accurately solves the problem of irregular grayscale difference changes of special vertical stripes under different exposure times and different line scanning cycles, and can well compensate for the grayscale changes of special vertical stripes, ensuring that the grayscale difference between columns and columns no longer changes under any exposure time and any line scanning cycle, thereby optimizing the image quality.

Description

Technical Field

[0001] The present invention belongs to the field of machine vision, and in particular relates to a method, system and medium for removing abnormal image pixels. Background Art

[0002] During the image quality detection stage of the line scan camera, it is often found that the image pixels are abnormal, that is, there are some vertical stripes in the image. The reason for the vertical stripes is usually that the photosensitive pixel rows of the line scan camera have abnormalities in individual pixels due to manufacturing process problems, which is manifested in the difference in grayscale values ​​between columns of the image. The common manifestation of vertical stripes is that the grayscale value difference between columns maintains a linear relationship with the change of grayscale value. In the prior art, grayscale compensation can be performed by vertical stripe correction algorithm, and linear fitting is performed with the grayscale value of a column of images as a unit, so as to achieve the purpose of making the grayscale mean values ​​of all columns in the image close or even the same under different light intensity conditions, thereby eliminating the vertical stripe phenomenon. However, there are some vertical stripes that do not simply change linearly with the change of grayscale value. It was found in the detection that the grayscale value of these special vertical stripes will change with the change of exposure time and line scanning cycle. The commonly used linear grayscale compensation method for this phenomenon cannot effectively eliminate the vertical stripe phenomenon under any exposure time and any line scanning cycle.

[0003] Therefore, in order to solve the above problems, the present invention provides a method, system and medium for removing abnormal image pixels. Summary of the invention

[0004] The purpose of the present invention is to overcome the above problems existing in the prior art and to provide a method, system and medium for removing abnormal pixel points in an image.

[0005] In order to achieve the above technical objectives and the above technical effects, the present invention is implemented through the following technical solutions:

[0006] A method for removing abnormal image pixels is provided, which obtains an original image formed by scanning a photosensitive pixel row in a dark field environment, and removes vertical lines in the original image. The removal method includes:

[0007] Confirm the position of the abnormal pixel points corresponding to the vertical stripes in the photosensitive pixel row, so as to filter out the abnormal pixel points corresponding to all the vertical stripes;

[0008] Divide into multiple exposure intervals, and correspond each exposure interval to a number of exposure times, so as to obtain a grayscale value corresponding to each exposure time;

[0009] Fitting each exposure time in all exposure intervals and its corresponding grayscale value to obtain compensation gain coefficients corresponding to all exposure intervals, so that all exposure times in the same exposure interval correspond to the same compensation gain coefficient;

[0010] Accumulate and calculate the interval compensation values ​​corresponding to all exposure intervals covered by the real-time exposure time and the maximum exposure time or the minimum exposure time to obtain the grayscale compensation value corresponding to the abnormal pixel point, thereby removing the vertical lines in the original image;

[0011] Among them, each abnormal pixel corresponds to multiple exposure intervals.

[0012] Furthermore, the line scanning period of the photosensitive pixel row is fixed.

[0013] Furthermore, the exposure times and their corresponding grayscale values ​​in all exposure intervals are fitted based on the least squares method.

[0014] Furthermore, the interval compensation value corresponding to the exposure interval is obtained by multiplying the exposure time variation by the compensation gain coefficient corresponding to the exposure interval.

[0015] Furthermore, the multiple exposure intervals corresponding to each abnormal pixel are symmetrically distributed.

[0016] Furthermore, the number of exposure times corresponding to each exposure interval is equal.

[0017] Furthermore, dividing the multiple exposure intervals includes: successively increasing the exposure time according to the same time difference and respectively counting the corresponding grayscale values ​​to calculate the slope value corresponding to each exposure time, thereby extracting several adjacent exposure times to determine the left and right boundaries of the exposure interval, wherein the difference between the slope values ​​corresponding to any two exposure times in the same exposure interval is less than the interval threshold.

[0018] Furthermore, confirming the position of the abnormal pixel corresponding to the vertical stripe in the photosensitive pixel row includes: extracting the row pixels corresponding to the photosensitive pixel row, and determining whether the difference between the grayscale value of each pixel and the average grayscale value of the row pixels exceeds the abnormal threshold, thereby screening out the pixel points corresponding to the difference exceeding the abnormal threshold as abnormal pixel points.

[0019] The present invention also provides an image pixel abnormality removal system, comprising:

[0020] The abnormal point screening module is used to confirm the position of the abnormal pixel points corresponding to the vertical lines in the photosensitive pixel row, so as to screen out the abnormal pixel points corresponding to all the vertical lines;

[0021] An interval division module is used to divide a plurality of exposure intervals, and each exposure interval corresponds to a number of exposure times, so as to obtain a grayscale value corresponding to each exposure time;

[0022] A coefficient analysis module is used to fit each exposure time in all exposure intervals and its corresponding grayscale value to obtain the compensation gain coefficients corresponding to all exposure intervals, so that all exposure times in the same exposure interval correspond to the same compensation gain coefficient;

[0023] The compensation analysis module is used to cumulatively calculate the interval compensation values ​​corresponding to all exposure intervals covered between the real-time exposure time and the maximum exposure time or the minimum exposure time, and obtain the grayscale compensation values ​​corresponding to the abnormal pixels, thereby removing the vertical lines in the original image.

[0024] The present invention also provides a computer-readable storage medium, comprising a computer program, wherein the computer program implements the above-mentioned removal method when executed by a processor.

[0025] The beneficial effects of the present invention are:

[0026] (1) The present invention divides the exposure intervals into multiple exposure intervals, and each exposure interval corresponds to a number of exposure times. When the grayscale value changes nonlinearly with the exposure time, the grayscale value of different exposure intervals can be compensated independently, and the compensation gain coefficients corresponding to all exposure intervals can be obtained by fitting. Therefore, any complex vertical stripe scene can be accurately and effectively removed. The interval compensation values ​​corresponding to all exposure intervals covered between the real-time exposure time and the maximum exposure time or the minimum exposure time are accumulated and calculated, and the independent compensation gain coefficients of each exposure interval are effectively utilized. The unit grayscale compensation value obtained by combining the exposure time change is accumulated to obtain the final grayscale compensation value. The problem of irregular grayscale difference changes of special vertical stripes under different exposure times and different line scanning cycles is accurately solved, and the grayscale change of special vertical stripes can be well compensated to ensure that the grayscale difference between columns does not change under any exposure time and any line scanning cycle, thereby optimizing the image quality.

[0027] (2) By determining that the difference between the slope values ​​corresponding to any two exposure times in the same exposure interval is less than the interval threshold, the present invention can effectively ensure that the grayscale values ​​corresponding to all exposure times in the same exposure interval have basically the same change trend, thereby providing a theoretical basis for subsequent calculations using the same compensation gain coefficient.

[0028] (3) The present invention can quickly screen and confirm the position of abnormal pixel points corresponding to vertical stripes in the photosensitive pixel row by screening out the pixel points corresponding to the difference exceeding the abnormal threshold as abnormal pixel points, thereby improving the overall operation efficiency and providing accurate data support for further optimizing the processing of special vertical columns. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0030] Figure 1 It is a flow chart of the removal method in the present invention;

[0031] Figure 2 It is a schematic diagram of the relationship between pixel value and exposure time in the present invention;

[0032] Figure 3 It is a schematic diagram of the compensation effect of the removal method in the present invention;

[0033] Figure 4 It is a structural block diagram of the removal system in the present invention. DETAILED DESCRIPTION

[0034] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0035] like Figure 1 As shown, this embodiment first provides a method for removing abnormal image pixels, obtains an original image formed by scanning a row of photosensitive pixels in a dark field environment, and removes vertical lines in the original image.

[0036] There are some special columns in the original image, and the grayscale mean of these columns will change with the exposure time and line scanning period in the dark field (excluding the influence of light). For normal columns, under no light conditions, the grayscale value only increases slightly with the increase of exposure time (dark current), and remains almost unchanged. However, the performance of special columns is different. Its change is similar to a quadratic function. It will increase first and then decrease, or decrease first and then increase with the increase of exposure time. Therefore, these columns need special column correction.

[0037] From the above, we can see that the grayscale values ​​of the pixels corresponding to these special vertical columns will change dynamically with the exposure time and line scanning cycle, such as Figure 2 As shown, the relationship between the grayscale value of pixel 543 and pixel 1410 and the exposure time is demonstrated. It can be found that:

[0038] When the exposure time is small and the exposure time is close to the line scan period, the grayscale value will change dramatically;

[0039] When the exposure time approaches 50% of the line scan period, its grayscale value begins to change slowly and eventually stabilizes.

[0040] In view of the above phenomenon, the removal method specifically includes the following steps:

[0041] The positions of abnormal pixels corresponding to the vertical stripes in the photosensitive pixel row are confirmed, thereby filtering out the abnormal pixels corresponding to all the vertical stripes.

[0042] As a specific embodiment of the present invention, confirming the position of abnormal pixel points corresponding to the vertical stripes in the photosensitive pixel row specifically includes the following steps: extracting the row pixels corresponding to the photosensitive pixel row, and determining whether the difference between the grayscale value of each pixel point and the average grayscale value of the row pixels exceeds the abnormal threshold, thereby screening out the pixel points corresponding to the difference exceeding the abnormal threshold as abnormal pixel points. For example, the abnormal threshold can be set to 2%. If the difference between the grayscale value of a certain pixel point and the average grayscale value of the row pixels is 5%, it is determined to be an abnormal pixel point. For the grayscale value of each pixel point, it can be either the average value of the vertical column pixels corresponding to the pixel point, or the pixel value of the corresponding pixel point position in the row pixel can be directly selected.

[0043] Divide into multiple exposure intervals, and correspond each exposure interval to a number of exposure times, so as to obtain a grayscale value corresponding to each exposure time; wherein each abnormal pixel corresponds to multiple exposure intervals.

[0044] As a specific embodiment of the present invention, dividing into multiple exposure intervals specifically includes the following steps: gradually increasing the exposure time according to the same time difference and respectively counting the corresponding grayscale values ​​to calculate the slope value corresponding to each exposure time, thereby extracting several adjacent exposure times to determine the left and right boundaries of the exposure interval, wherein the difference between the slope values ​​corresponding to any two exposure times in the same exposure interval is less than the interval threshold.

[0045] Since the difference between the slope values ​​corresponding to any two exposure times in the same exposure interval is smaller than the interval threshold, the grayscale values ​​corresponding to all exposure times in the same exposure interval can maintain basically the same trend, thereby providing a basis for adopting the same compensation gain coefficient.

[0046] The exposure times and their corresponding grayscale values ​​in all exposure intervals are fitted to obtain the compensation gain coefficients corresponding to all exposure intervals, so that all exposure times in the same exposure interval correspond to the same compensation gain coefficient.

[0047] As a specific implementation of the present invention, all exposure times and their corresponding grayscale values ​​obtained statistically in each exposure interval can be fitted based on the least squares method, so as to obtain the fitting coefficient corresponding to each exposure interval according to the fitting result, that is, the compensation gain coefficient. In terms of overall logic, a piecewise linear function is used to fit the relationship curve between the grayscale value of the pixel point and the exposure time, and the transformation relationship between the exposure time and the grayscale value is mapped to a linear function, and finally the difference is dynamically compensated according to the exposure time at the current sampling moment.

[0048] The interval compensation values ​​corresponding to all exposure intervals covered between the real-time exposure time and the maximum exposure time or the minimum exposure time are accumulated and calculated to obtain the grayscale compensation values ​​corresponding to the abnormal pixels, thereby removing the vertical lines in the original image.

[0049] As can be seen from the above, the product of the compensation gain coefficient and the change in exposure time is the corresponding unit gray compensation value. That is, the interval compensation value corresponding to the exposure interval can be obtained by multiplying the change in exposure time and the compensation gain coefficient of the corresponding exposure interval. Since different exposure intervals have different compensation gain coefficients, the total change in exposure time covered between any real-time exposure time and the maximum or minimum exposure time corresponds to multiple exposure intervals. Then, taking different exposure intervals as the change in exposure time and combining the corresponding compensation gain coefficients, the unit gray compensation values of each segment are obtained respectively, and the sum of all unit gray compensation values can be used to obtain the total gray compensation value.

[0050] The specific operation details are as follows:

[0051] Since not all vertical columns are special vertical columns where the gray value changes dynamically with the exposure time and the line scan period, in order to further optimize the application scenario of the above removal method, as a preferred implementation, in the dark field case, the line scan period can be fixed at Prd, and then the exposure time Exp (0 < Exp < Prd) is continuously adjusted and increased, and the change in the gray value of the vertical column is recorded. The vertical columns with the gray value changing first increasing and then decreasing or first decreasing and then increasing are marked. These vertical columns can be considered as special vertical columns, and then only the above removal method is used for this special vertical column, while for the vertical stripes where the gray value difference between columns remains linearly related with the change in the gray value, linear fitting can be directly used for gray compensation.

[0052] As can be seen from the above, the change of the special vertical column is similar to a quadratic function, which will increase first and then decrease or decrease first and then increase with the increase of the exposure time. Therefore, in view of the above characteristics of the special vertical column, as a preferred implementation, multiple exposure intervals corresponding to each abnormal pixel point can be symmetrically distributed. At the same time, in order to further ensure that each exposure interval has a sufficient statistical quantity and facilitate subsequent operations, the number of exposure times corresponding to each exposure interval can be made equal. To simplify the variables, the line scan period of the photosensitive pixel row can be fixed.

[0053] Specifically, after confirming the position of the abnormal pixel points corresponding to the vertical stripes in the photosensitive pixel row, with the line scan period Prd remaining unchanged, the exposure time Exp is divided into N intervals from 0 to Prd. Since the gray value change trend of the special vertical column is "symmetric", the exposure interval division also needs to be symmetric. That is, N / 2 intervals are divided within the exposure time from 0 to Prd / 2, and N / 2 intervals are divided from Prd / 2 to Prd. Each interval is evenly divided into M segments. At this time, the exposure time segmentation point is E ij (0 ≤ i ≤ N - 1, 0 ≤ j ≤ M - 1). Then adjust the exposure time Exp = E ij , and record the gray value G of the special vertical column ij(0≤i≤N-1, 0≤j≤M-1), linear fitting is performed for the gray value and exposure time changes in each interval to obtain the compensation gain coefficient P i (0≤i≤N-1), the specific calculation process is as follows:

[0054]

[0055] At this time P i is the compensation gain coefficient corresponding to each exposure interval when the line scanning period is Prd, temp i 、res i , Rau i , Q i All participated in P i The default parameters for the operation.

[0056] Obtain the compensation gain coefficient P for each exposure interval i After that, we need to determine the relationship between exposure time and line scanning period to ensure that when line scanning period and exposure time change at the same time, we can select the appropriate compensation gain coefficient P. i , calculate the grayscale compensation value Bix. Taking N=6, i.e. 6 exposure intervals, as an example, the specific calculation process is as follows:

[0057]

[0058]

[0059] Among them, T, id, E 0 、E 1 、E 2 、E 3 、E 4 、E 5 are the preset parameters for Bix calculation, P id , P id+1 , P id+2 are the compensation gain coefficients corresponding to the exposure range.

[0060] The following is a further explanation with specific cases and effects:

[0061] like Figure 3 As shown in the figure, the dynamic compensation effect is demonstrated when the line scan period is 400us and the number of exposure intervals is 6. The interval division is shown in the following table. It can be seen that the output response difference of the pixel is reduced from 20DN to 3DN, which can ensure the consistency of image quality.

[0062] Table 1 Exposure interval-exposure schedule

[0063]

[0064] like Figure 4 As shown, the present invention also provides an image pixel abnormality removal system, comprising:

[0065] The abnormal point screening module is used to confirm the position of the abnormal pixel points corresponding to the vertical lines in the photosensitive pixel row, so as to screen out the abnormal pixel points corresponding to all the vertical lines;

[0066] An interval division module is used to divide a plurality of exposure intervals, and each exposure interval corresponds to a number of exposure times, so as to obtain a grayscale value corresponding to each exposure time;

[0067] A coefficient analysis module is used to fit each exposure time in all exposure intervals and its corresponding grayscale value to obtain the compensation gain coefficients corresponding to all exposure intervals, so that all exposure times in the same exposure interval correspond to the same compensation gain coefficient;

[0068] The compensation analysis module is used to cumulatively calculate the interval compensation values ​​corresponding to all exposure intervals covered between the real-time exposure time and the maximum exposure time or the minimum exposure time, and obtain the grayscale compensation values ​​corresponding to the abnormal pixels, thereby removing the vertical lines in the original image.

[0069] The specific operation methods of the abnormal point screening module, the interval division module, the coefficient analysis module and the compensation analysis module can refer to the above removal method for operation.

[0070] The third aspect of the present invention further provides a computer-readable storage medium, comprising a computer program, wherein the computer program implements the above-mentioned removal method when executed by a processor.

[0071] In practical applications, the computer-readable storage medium may adopt any combination of one or more computer-readable media. The computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium may be, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, device or device.

[0072] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, which carry computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

Claims

1. A method for removing abnormal image pixels, obtaining an original image formed by scanning a photosensitive pixel row in a dark field environment, and removing vertical lines in the original image, characterized in that: Removal methods include: Confirm the position of abnormal pixel points corresponding to the vertical stripes in the photosensitive pixel row, so as to filter out the abnormal pixel points corresponding to all the vertical stripes; Divide into multiple exposure intervals, and correspond each exposure interval to a number of exposure times, so as to obtain a grayscale value corresponding to each exposure time; Fitting each exposure time in all exposure intervals and its corresponding grayscale value to obtain compensation gain coefficients corresponding to all exposure intervals, so that all exposure times in the same exposure interval correspond to the same compensation gain coefficient; Accumulate and calculate the interval compensation values ​​corresponding to all exposure intervals covered by the real-time exposure time and the maximum exposure time or the minimum exposure time to obtain the grayscale compensation value corresponding to the abnormal pixel point, thereby removing the vertical lines in the original image; Among them, each abnormal pixel corresponds to multiple exposure intervals.

2. The method for removing abnormal image pixels according to claim 1, characterized in that: The line scanning period of the photosensitive pixel row is fixed.

3. The method for removing abnormal image pixels according to claim 1, characterized in that: The exposure time and its corresponding grayscale value in all exposure intervals are fitted based on the least squares method.

4. The method for removing abnormal image pixels according to claim 1, characterized in that: The interval compensation value corresponding to the exposure interval is obtained by multiplying the exposure time variation and the compensation gain coefficient corresponding to the exposure interval.

5. The method for removing abnormal image pixels according to any one of claims 1 to 4, characterized in that: The multiple exposure intervals corresponding to each abnormal pixel are symmetrically distributed.

6. The method for removing abnormal image pixels according to claim 5, characterized in that: The number of exposure times corresponding to each exposure interval is equal.

7. The method for removing abnormal image pixels according to any one of claims 1 to 4, characterized in that: Dividing multiple exposure intervals includes: successively increasing the exposure time according to the same time difference and counting the corresponding grayscale values ​​respectively to calculate the slope value corresponding to each exposure time, thereby extracting several adjacent exposure times to determine the left and right boundaries of the exposure interval, wherein the difference between the slope values ​​corresponding to any two exposure times in the same exposure interval is less than the interval threshold.

8. The method for removing abnormal image pixels according to any one of claims 1 to 4, characterized in that: Confirming the position of the abnormal pixel point corresponding to the vertical stripe in the photosensitive pixel row includes: extracting the row pixels corresponding to the photosensitive pixel row, and determining whether the difference between the grayscale value of each pixel point and the average grayscale value of the row pixels exceeds the abnormal threshold, thereby screening out the pixel points corresponding to the difference exceeding the abnormal threshold as abnormal pixel points.

9. An image pixel abnormality removal system, characterized in that: include: The abnormal point screening module is used to confirm the position of the abnormal pixel points corresponding to the vertical lines in the photosensitive pixel row, so as to screen out the abnormal pixel points corresponding to all the vertical lines; An interval division module is used to divide a plurality of exposure intervals, and each exposure interval corresponds to a number of exposure times, so as to obtain a grayscale value corresponding to each exposure time; A coefficient analysis module is used to fit each exposure time in all exposure intervals and its corresponding grayscale value to obtain the compensation gain coefficients corresponding to all exposure intervals, so that all exposure times in the same exposure interval correspond to the same compensation gain coefficient; The compensation analysis module is used to cumulatively calculate the interval compensation values ​​corresponding to all exposure intervals covered between the real-time exposure time and the maximum exposure time or the minimum exposure time, and obtain the grayscale compensation values ​​corresponding to the abnormal pixels, thereby removing the vertical lines in the original image.

10. A computer-readable storage medium comprising a computer program, characterized in that: When the computer program is executed by a processor, the removal method according to any one of claims 1 to 8 is implemented.

Images