A non-uniformity correction method based on an image intensifier non-uniformity test system

Abstract

The application discloses a non-uniformity correction method based on an image intensifier non-uniformity test system, and mainly solves the technical problem that the existing non-uniformity correction method is easy to introduce systematic statistical deviation, and finally leads to the distortion of the gain and bias correction coefficient solution of the acquisition equipment. The application comprises the following steps: building an image intensifier non-uniformity test system; obtaining flat field images and dark field images under different gray test points; image superposition and average processing; calculating a gray self-adaptive weight, calculating the global average gray value of the whole frame image according to the gray self-adaptive weight, and obtaining the mapping relationship of the group uniform image and the flat field image; constructing a non-uniformity correction formula by using the least square method; collecting the image at the output end of the image intensifier, performing non-linear correction, and then calculating the non-uniformity, and finally completing the image data output. The application has excellent wide dynamic brightness adaptation performance, and can meet the imaging acquisition and detection test requirements in different brightness gradient scenes.

Description

Technical Field

[0001] This invention relates to a non-uniformity correction method, specifically a non-uniformity correction method based on an image intensifier non-uniformity testing system. Background Technology

[0002] Image intensifiers are core photoelectric conversion and brightness amplification devices in low-light night vision, photoelectric detection, low-light imaging, and high-speed transient optical testing systems. They are widely used in military and civilian fields such as individual soldier night vision observation and aiming, vehicle-mounted and airborne photoelectric pods, border security monitoring, astronomical low-light detection, biomedical low-light imaging, and transient optical diagnostics. With the iteration of third-generation, fourth-generation, and super-second-generation image intensifiers, and the development of low-light imaging systems from manual visual observation to digital intelligent target recognition and image quantization measurement, the industry has imposed stringent requirements on the non-uniformity of the image output from image intensifiers. Non-uniformity correction technology for acquisition systems has become a core technological necessity in the research, testing, and application of low-light devices.

[0003] Chinese patent CN110855915A discloses a "CCD camera response non-uniformity correction method and device for achieving dark current compensation". This method establishes a pixel-level dark current compensation model by acquiring the dark field response of the CCD, and combines the response correction coefficients obtained by flat field calibration to perform pixel-by-pixel joint compensation on the acquired image to achieve the correction of CCD camera response non-uniformity. However, the dark current model relies on offline calibration and is difficult to adapt to dynamic changes. In actual imaging systems, factors such as spatial non-uniformity of the light source and vignetting of the optical lens also affect the non-uniformity.

[0004] Chinese patent CN121678126A discloses a "High-precision detection system and testing method for the uniformity of output images from a low-light image intensifier." This method utilizes an adjustable aperture to define the instantaneous spatial aperture involved in the measurement. A single high-linearity, high-sensitivity detector sequentially integrates the local brightness at various positions along the horizontal diameter during scanning, effectively mapping the originally two-dimensional distributed area array image information into a one-dimensional brightness distribution curve with respect to position. However, the reciprocating motion of the translation stage during scanning introduces accumulated errors, resulting in missed detection points. Furthermore, the testing system lacks non-uniformity correction, leading to significant errors in detecting image intensification non-uniformity. Therefore, an algorithm suitable for a wide brightness range and requiring high-precision non-uniformity correction is needed.

[0005] When performing non-uniformity correction on CCD / CMOS acquisition devices, due to the lens vignetting effect, images acquired under uniform light source illumination exhibit a brightness distribution characteristic of being bright in the center and dark around the edges. Figure 1As shown. If the arithmetic mean is used directly to calculate the reference gray level of the entire field, a large number of high gray level pixels in the central region will dominate the statistical results, causing the calculated average gray level to be higher than the actual irradiance gray level corresponding to the uniform light source of the integrating sphere. This introduces a systematic statistical bias, which ultimately leads to distortion in the calculation of the gain and offset correction coefficients of the acquisition equipment. Summary of the Invention

[0006] The purpose of this invention is to solve the technical problem that existing non-uniformity correction methods are prone to introducing systematic statistical biases, which ultimately lead to distortion in the calculation of acquisition device gain and bias correction coefficients. Instead, this invention provides a non-uniformity correction method based on an image intensifier non-uniformity testing system.

[0007] To achieve the above objectives, the technical solution provided by this invention is as follows:

[0008] A non-uniformity correction method based on an image intensifier non-uniformity testing system, characterized by the following steps:

[0009] Step 1: Set up an image intensifier non-uniformity testing system

[0010] The image intensifier non-uniformity testing system includes an image intensifier, a data acquisition device, a translation stage 1, a translation stage 2, and a uniform light source. Translation stages 1 and 2 are placed on one side of the uniform light source, and the image intensifier and data acquisition device are placed on translation stages 1 and 2, respectively. The image intensifier is moved to a designated position on one side of the uniform light source using translation stage 1. A target is set up on the radiation surface of the uniform light source, and translation stage 2 is adjusted so that the data acquisition device can clearly image the target.

[0011] Step 2: Set up equally spaced grayscale test points. Groups, each group indivual, , Adjust the brightness of the uniform light source to the level of the acquisition device. Group grayscale test points, collect and store them separately. The system acquires a set of flat-field images. Simultaneously, after each flat-field image acquisition, the uniform light source is turned off and the optical path is blocked, while keeping other system parameters constant. The images are then acquired and stored synchronously. Group of dark field images;

[0012] Step 3: Perform stacking averaging on the flat field image and the dark field image respectively to obtain the stacked average flat field image and the dark field image. Then, subtract the dark field image from the stacked average flat field image to obtain the image to be corrected. Group of images;

[0013] Step 4, for For a group of images, iterate through the grayscale values ​​of each pixel, filter out the grayscale interval corresponding to the highest peak of the histogram, and take the average of the grayscale values ​​of all pixels within the interval as the benchmark. and based on this benchmark Calculate the grayscale adaptive weights to obtain Group weighting chart;

[0014] Step 5: Calculate the global average grayscale value for each weighted image, generate a constant-size matrix, and uniformly fill all pixels in the matrix with this average grayscale value to obtain the result. Group uniform images, and Group uniform image and The pixel-by-pixel mapping relationship of the flat-field image;

[0015] Step 6: Based on the above mapping relationship, construct the least squares cost function with the principle of minimizing the sum of squared residuals, and use the least squares method to fit and solve the intrinsic gain and dark base offset of each pixel to obtain the non-uniformity correction formula.

[0016] Step 7: Move the No. 1 translation stage so that the input end of the image intensifier faces the radiation surface of the uniform light source. Under standard operating conditions, the acquisition device uses the image output end of the image intensifier to acquire multiple frames of flat field images under standard operating conditions to be corrected. After performing non-uniformity correction according to the non-uniformity correction formula, the non-uniformity is quantitatively solved to complete the non-uniformity correction.

[0017] Furthermore, the feature is that, in step 3, the flat field image and the dark field image are respectively subjected to superposition averaging processing using the following formula to obtain the flat field image after superposition averaging processing. and dark field images :

[0018] ;

[0019] In the formula: For the same group Zhang Ping's field images are overlaid; For the same group Overlay of dark field images; and These are the horizontal and vertical coordinates of the pixel; ;

[0020] The following formula is used to subtract the dark field image from the flat field image after superposition and averaging to obtain the image to be corrected. Group of images :

[0021] - .

[0022] Furthermore, the feature is that, in step 4, the following formula is used, based on the benchmark. Calculate the grayscale adaptive weights to obtain the weight map. :

[0023] ;

[0024] In the formula: To adjust the intensity coefficient, 0 < <1.

[0025] Furthermore, the feature is that, in step 5, the global average gray value of the weighted image is calculated using the following formula. :

[0026] ;

[0027] In the formula: The width of the weighted map is represented by the number of pixels. This represents the height of the weighted map in pixels.

[0028] Furthermore, the feature is that, in step 6, the least squares cost function for:

[0029] ;

[0030] In the formula: For the first Group of uniform images;

[0031] The non-uniformity correction formula is as follows:

[0032] ;

[0033] In the formula: This represents the average grayscale value after image non-uniformity correction. This is the inherent gain; For dark base offset; This is a flat-field image.

[0034] Furthermore, in step 7, statistical analysis is performed on the global grayscale data of the flat field image to be corrected. After non-uniformity correction is performed according to the non-uniformity correction formula, the standard deviation of the image grayscale is calculated. Average gray level of the entire image Based on the following non-uniformity standard calculation formula, the spatial non-uniformity of the image output image of the image intensifier is calculated. Quantitative solution:

[0035] .

[0036] Compared with the prior art, the present invention has the following beneficial technical effects:

[0037] This invention corrects the brightness distribution of the image by adaptive weight allocation of grayscale, balances the global pixel grayscale performance, and performs non-uniformity correction within a wide dynamic range of brightness. It has excellent wide dynamic range brightness adaptation performance and can meet the imaging acquisition and detection test requirements under different brightness gradient scenarios. Attached Figure Description

[0038] Figure 1 The original images are from the existing acquisition system.

[0039] Figure 2 This is a schematic diagram of the image intensifier non-uniformity testing system in an embodiment of the present invention;

[0040] Figure 3 This is a flowchart of an embodiment of the present invention;

[0041] Figure 4 The image shows the correction effect of existing non-uniformity correction methods.

[0042] Figure 5 This is a diagram illustrating the correction effect of an embodiment of the present invention.

[0043] The attached figures are labeled as follows:

[0044] 1-Image intensifier, 2-Acquisition device, 3-Translation stage 1, 4-Translation stage 2, 5-Uniform light source, 6-Power supply, 7-PC. Detailed Implementation

[0045] To make the objectives, advantages, and features of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Those skilled in the art should understand that these embodiments are merely used to explain the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0046] like Figure 3 As shown in the figure, this embodiment provides a non-uniformity correction method based on an image intensifier non-uniformity testing system, which specifically includes the following steps:

[0047] Step 1, such as Figure 2As shown, the image intensifier non-uniformity testing system in this embodiment includes an image intensifier 1, a data acquisition device 2, a translation stage 1 3, a translation stage 2 4, and a uniform light source 5. The translation stages 1 3 and 2 4 are placed on one side of the uniform light source 5, and the image intensifier 1 and the data acquisition device 2 are placed on the translation stages 1 3 and 2 4, respectively. The image intensifier 1 is powered by the power supply 6. The image intensifier 1 is translated to a designated position on one side of the uniform light source 5 by the translation stage 1 3. A target is set up on the radiation surface of the uniform light source 5. The translation stage 2 4 is adjusted so that the data acquisition device 2 can clearly image the image. The data acquisition device 2 is then connected to the PC 7.

[0048] Step 2: To achieve accurate non-uniformity correction across the entire brightness range, set up equally spaced grayscale test points. Groups, each group Adjust the brightness of each uniform light source to [value]. Group test points, collect and store data respectively The system acquires a set of flat-field images; simultaneously, after each flat-field image acquisition, the light source is turned off and the optical path is blocked, while keeping the device parameters constant, and the images are acquired synchronously. A set of dark field images were used to complete the acquisition of raw image data under multi-gradient brightness conditions.

[0049] Step 3: Image preprocessing to remove noise from the dark substrate, resulting in a flat-field image. Dark field image is The flat-field image and the dark-field image are averaged separately according to Formula 1 to smooth random noise and obtain a stable basis:

[0050] (1)

[0051] In the formula For the same group Zhang Ping's field images are overlaid; For the same group Overlay of dark field images; and These are the horizontal and vertical coordinates of the pixel; For the first Group flat field images; For the first Group of dark field images; .

[0052] According to Formula 2, the image to be corrected is obtained by subtracting the dark field image from the processed flat field image. Group of images :

[0053] - (2)

[0054] Step 4: Traverse the grayscale values ​​of pixels and filter for the highest peak in the histogram (the grayscale range corresponding to a large, normal, and uniform area [R]). L ,R H The average grayscale value of all pixels within the interval is taken as the benchmark. The weight map is obtained by calculating the grayscale adaptive weights according to Formula 3, where... To standardize the intensity coefficient, 0 < <1, determined based on the actual brightness of the image acquired by the acquisition device. Value magnitude, with relatively small differences in brightness. Taking smaller values ​​(0.3~0.5) results in greater differences in brightness. Take a larger value (0.7~0.9);

[0055] (3)

[0056] Step 5: Calculate the global average gray value of the weighted map according to Formula 4. Generate a constant-size matrix, and fill all pixels in the matrix with the global average gray value. ,get Group uniform image ,as well as Group uniform image and The mapping relationship of each pixel in the flat field image.

[0057] (4)

[0058] In the formula: The width of the weighted map is represented by the number of pixels. This represents the height of the weighted map in pixels.

[0059] Step Six: Based on the mapping relationship described above, construct the least squares cost function according to Formula 5 to minimize the sum of squared residuals, and then use the least squares method to fit and solve for the intrinsic gain of each pixel pixel by pixel. offset from dark base Formula 6 for non-uniformity correction is obtained. At this point, the wide brightness range and high precision non-uniformity correction of the acquisition device is completed. Subsequent non-uniformity detection of the image intensifier does not require successive correction. A periodic correction scheme can be used to stably maintain the test and measurement accuracy.

[0060] (5)

[0061] (6)

[0062] In the formula: This represents the average gray value after image non-uniformity correction.

[0063] Step 7: Move translation stage 1 so that the input of the image intensifier faces the uniform light source radiation surface. The acquisition device uses the image from the output of the image intensifier. The acquisition device performs image acquisition on the fluorescent screen at the output of the image intensifier under standard operating conditions. Under the test conditions of ensuring uniform light source, fixed system parameters, and stable operating voltage and gain, multiple frames of flat field images to be corrected under uniform illumination are acquired. By performing statistical analysis on the global grayscale data of the acquired image, and after correcting for the non-uniformity of the acquisition device according to Formula 6, the standard deviation of the image grayscale is calculated. Average gray level of the entire image Based on the non-uniformity standard calculation formula 7, the spatial non-uniformity of the image output image of the image intensifier is quantitatively solved, thereby characterizing the consistency level of the brightness distribution of the image intensifier across the entire screen and objectively evaluating the non-uniformity characteristics of the image intensifier.

[0064] (7)

[0065] See Figure 4 and Figure 5 Using a 12-bit CCD camera, images were acquired at 10 different brightness levels. Ten sets of grayscale values ​​for a single pixel were selected, and the non-uniformity correction formula was calculated using both conventional non-uniformity correction and the method of this invention. Comparison shows that conventional non-uniformity correction results in a loosely distributed point distribution and large residuals, indicating a generally weak linear correlation between the grayscale value to be corrected and the corrected grayscale value. The method of this invention, on the other hand, exhibits a compact distribution, better fitting results, and smaller correction error.

[0066] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the present invention.

Claims

1. A non-uniformity correction method based on an image intensifier non-uniformity testing system, characterized in that, Includes the following steps: Step 1: Build an image intensifier non-uniformity testing system; The image intensifier non-uniformity testing system includes an image intensifier, a data acquisition device, a translation stage 1, a translation stage 2, and a uniform light source. Translation stages 1 and 2 are placed on one side of the uniform light source, and the image intensifier and data acquisition device are placed on translation stages 1 and 2, respectively. The image intensifier is moved to a designated position on one side of the uniform light source using translation stage 1. A target is set up on the radiation surface of the uniform light source, and translation stage 2 is adjusted so that the data acquisition device can clearly image the target. Step 2: Set up equally spaced grayscale test points. Groups, each group indivual, , Adjust the brightness of the uniform light source to... Group grayscale test points, collect and store them separately. The system acquires a set of flat-field images. Simultaneously, after each flat-field image acquisition, the uniform light source is turned off and the optical path is blocked, while keeping other system parameters constant. The images are then acquired and stored synchronously. Group of dark-field images; Step 3: Perform stacking averaging on the flat field image and the dark field image respectively to obtain the stacked average flat field image and the dark field image. Then, subtract the dark field image from the stacked average flat field image to obtain the image to be corrected. Group of images; Step 4, for For a group of images, iterate through the grayscale values ​​of each pixel, filter out the grayscale interval corresponding to the highest peak of the histogram, and take the average of the grayscale values ​​of all pixels within the interval as the benchmark. and based on this benchmark Calculate the grayscale adaptive weights to obtain Group weighting chart; Step 5: Calculate the global average grayscale value for each weighted image, generate a constant-size matrix, and uniformly fill all pixels in the matrix with the global average grayscale value to obtain the result. Group uniform images, and Group uniform image and The pixel-by-pixel mapping relationship of the flat-field image; Step 6: Based on the above mapping relationship, construct the least squares cost function with the principle of minimizing the sum of squared residuals, and use the least squares method to fit and solve the intrinsic gain and dark base offset of each pixel to obtain the non-uniformity correction formula. Step 7: Move the No. 1 translation stage so that the input end of the image intensifier faces the radiation surface of the uniform light source. The acquisition device acquires the image output end of the image intensifier under standard working conditions, obtains multiple frames of flat field images to be corrected under standard working conditions, performs non-uniformity correction on it according to the non-uniformity correction formula, and then quantitatively solves the non-uniformity to complete the non-uniformity correction.

2. The non-uniformity correction method based on an image intensifier non-uniformity testing system according to claim 1, characterized in that, In step 3, the flat field image and the dark field image are respectively subjected to stacking and averaging processing using the following formula to obtain the flat field image after stacking and averaging. and dark field images : ； In the formula: For the same group Zhang Ping's field images are overlaid; For the same group Overlay of dark field images; and These are the horizontal and vertical coordinates of the pixel; ; The following formula is used to subtract the dark field image from the flat field image after superposition and averaging to obtain the image to be corrected. Group of images : - 。 3. The non-uniformity correction method based on an image intensifier non-uniformity testing system according to claim 2, characterized in that, In step 4, the following formula is used, based on the benchmark. Calculate the grayscale adaptive weights to obtain the weight map. : ； In the formula: To adjust the intensity coefficient, 0 < <1.

4. The non-uniformity correction method based on an image intensifier non-uniformity testing system according to claim 3, characterized in that, In step 5, the global average gray value of the weighted image is calculated using the following formula. : ； In the formula: The width of the weighted map is represented by the number of pixels. This represents the height of the weighted map in pixels.

5. A non-uniformity correction method based on an image intensifier non-uniformity testing system according to any one of claims 2-4, characterized in that, In step 6, the least squares cost function for: ； In the formula: For the first Group of uniform images; The non-uniformity correction formula is as follows: ； In the formula: This represents the average grayscale value after image non-uniformity correction. This is the inherent gain; For dark base offset; This is a flat-field image.

6. The non-uniformity correction method based on an image intensifier non-uniformity testing system according to claim 5, characterized in that, In step 7, statistical analysis is performed on the global grayscale data of the flat field image to be corrected. After non-uniformity correction is performed according to the non-uniformity correction formula, the standard deviation of the image grayscale is calculated. Average gray level of the entire image Based on the following non-uniformity standard calculation formula, the non-uniformity of the image output image of the image intensifier is calculated. Quantitative solution: 。

Images