A method for joint correction of dark field and radiometric response non-uniformity of a color area array camera

Abstract

The application discloses a dark field and radiation response non-uniformity joint correction method of a color area array camera. The dark field and radiation response non-uniformity correction coefficients of each channel of the camera under different temperature environments and under a uniform light environment are collected; the original output Bayer image of the camera is acquired under a certain environmental temperature and a certain exposure time; the theoretical dark field image data corresponding to each channel in the Bayer image is calculated according to the environmental temperature and the exposure time of the Bayer image and the dark field non-uniformity correction coefficient; after the difference between the image to be corrected and the theoretical dark field image is calculated, the radiation response non-uniformity correction is performed. After the corrected image data and the background bias signal are added, the image after the dark field and radiation response non-uniformity joint correction is obtained. The method can eliminate the influence of the internal dark current non-uniformity and the radiation response non-uniformity of each channel of the color area array camera under different temperatures on the output image.

Description

Technical Field

[0001] This invention relates to the field of imaging with color area array cameras, and more particularly to a method for joint correction of dark field and radiation response inhomogeneity in color area array cameras. Background Technology

[0002] Cameras are increasingly widely used in people's daily lives, scientific research, military, biomedicine, and aerospace. Color cameras play an irreplaceable role because they can reflect the color information of the photographed target. The output image of a color camera can be considered to be composed of images from three color channels: R, G, and B. Differences in photoelectric conversion efficiency at the front end of the detector and inconsistencies in the back end drive signals and processing circuits result in the output image of each channel of a color camera not being an ideally uniform image. At the same time, differences in photoelectric conversion efficiency between channels also lead to color differences between the color information output by the color camera and the true color of the photographed target. Therefore, the non-uniformity in a color camera includes both the non-uniformity within each channel and the inconsistency between the color response and the true color components between channels. To more accurately correct the color image output by the camera, it is necessary to first perform non-uniformity correction on each channel of the color camera.

[0003] Conventional non-uniformity correction methods include single-point, two-point, multi-point, and scene-based correction methods, but they do not consider the influence of temperature on image non-uniformity and have poor adaptability to different ambient temperatures. Summary of the Invention

[0004] The purpose of this invention is to propose a non-uniformity correction method for color area array cameras with stronger temperature adaptability by jointly correcting the non-uniformity of dark field and radiation response of images from each channel of a color camera.

[0005] Unlike conventional non-uniformity correction methods, this method comprehensively considers the impact of non-uniformity of dark current growth between pixels and non-uniformity of radiometric response on the output image within each channel of a color area array camera. It employs a correction algorithm that first performs single-sided dark-field correction on the Bayer image based on ambient temperature and exposure time, then performs bilateral radiometric response non-uniformity correction within each channel, and finally performs inverse linear reconstruction of the dark-field signal. This method is widely applicable to non-uniformity correction of color camera output images under different ambient temperatures, exposure times, and lighting conditions. The technical solution of this invention is as follows:

[0006] A method for joint correction of dark field and radiometric response inhomogeneity of a color area array camera, the specific steps of which are as follows:

[0007] (1) Place the camera in a dark room, close the protective cover in front of the camera lens, power on the camera, and put the camera in a dark environment.

[0008] (2) Set the ambient temperature of the camera to T and stabilize for 60 minutes;

[0009] (3) The minimum exposure time that the camera can be set to is t darkMin The maximum exposure time is t darkMax ; in t darkMin With t darkMax At least m exposure time points t are selected evenly between them. darkExp And acquire n dark field images at each exposure time point; extract dark field images of each color channel R, G, B at all exposure time points according to the distribution of color channels in the detector;

[0010] (4) In color channel R, at exposure time point t darkExp Let X be the average grayscale value of the pixel in the i-th row and j-th column of the dark field mean image at each exposure time, expressed as Dark. R By plotting Y and performing a linear fit according to the following formula, the dark field non-uniformity correction coefficient k of the current pixel can be obtained. R_dark With b R_dark ;

[0011] Dark R (i,j)=k R_dark (i,j)*t darkExp +b R_dark (i,j)

[0012] (5) Change the pixel coordinates to obtain the dark field non-uniformity correction coefficients for all pixels in the current color channel.

[0013] (6) Repeat steps (4)-(5) to obtain the dark field non-uniformity correction coefficients of all pixels in color channel G and color channel B.

[0014] (7) Change the ambient temperature and repeat steps (2)-(6) to calculate the dark field non-uniformity correction coefficients for each color channel under all ambient temperatures.

[0015] (8) With temperature T as X, the coefficient K of a single pixel at each temperature point dark Let Y be the dark current coefficients a and b. After performing high-order curve fitting according to the following formula, e is the natural constant.

[0016] K dark (i,j)=a*e b*T

[0017] (9) Turn on the integrating sphere, set the operating current, and bring the camera and integrating sphere to a stable operating state; in a uniformly illuminated environment with the integrating sphere as the sole light source, obtain the camera's minimum exposure time t. brightMin With maximum exposure time tbrightMax ;

[0018] (10) At the minimum exposure time t brightMin With maximum exposure time t brightMax An average of m exposure time points are set between each exposure time point. At each exposure time point, n bright and dark field images are collected, and the corresponding mean image is calculated.

[0019] (11) Subtract the mean image of the bright field from the mean image of the dark field to obtain the difference image at a single exposure time point;

[0020] (12) After linear fitting with the exposure time as X and the gray value of each pixel in the difference image at different exposure time points as Y, the non-uniformity correction coefficients k(i,j) and b(i,j) of the radiation response of each pixel in the image are obtained.

[0021] (13) According to the color channel distribution in the detector, extract the radiation response non-uniformity correction coefficients k, b and kAve, bAve for each channel;

[0022] (14) Obtain a raw output Bayer image from the camera at a certain temperature environment T0 and exposure time t0; Substitute the current temperature environment T0 and exposure time t0 into the formulas in steps (4) and (8) to calculate the current theoretical dark field Bayer image.

[0023] (15) Subtract the original output Bayer image from the theoretical dark field Bayer image to obtain the difference image. diff After processing the detector array into channels, the non-uniformity of the radiation response of each pixel in each channel is corrected according to the following formula to obtain the image after non-uniformity correction of radiation response.

[0024]

[0025] (16) Add the image after radiation response non-uniformity correction to the background bias of the corresponding pixel to obtain the final Bayer image after joint correction of dark field and radiation response non-uniformity.

[0026] image Modify =imagediff Modify +b darkAve .

[0027] The beneficial effects of this invention are:

[0028] This invention employs a correction algorithm that first performs single-sided dark-field correction on Bayer images based on ambient temperature and exposure time, then performs bilateral radiative response non-uniformity correction within each channel, and finally performs inverse linear reconstruction of the dark-field signal. This algorithm is widely applicable to the non-uniformity correction of color camera output images under different ambient temperatures, exposure times, and lighting conditions. This method provides a new approach to non-uniformity correction for color cameras and also provides a theoretical basis for more accurate color correction of cameras in a wide-temperature environment. Attached Figure Description

[0029] Figure 1 This is a flowchart of the joint correction process for dark field and radiation response non-uniformity in this invention. Detailed Implementation

[0030] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0031] This invention relates to the field of imaging with color area array cameras, and more particularly to a method for joint correction of dark field and radiation response inhomogeneity in color area array cameras.

[0032] This invention can be roughly divided into four steps: 1) determining the dark field non-uniformity correction coefficients; 2) determining the higher-order function coefficients of dark current; 3) determining the radiation response non-uniformity correction coefficients; 4) performing joint non-uniformity correction based on environmental conditions, camera exposure time, and various correction coefficients; such as Figure 1 As shown,

[0033] Step 1: Determining the dark field non-uniformity correction coefficient.

[0034] 1.1 Place the camera in a darkroom, close the protective cover in front of the camera lens, power on the camera, and put the camera in a dark environment;

[0035] 1.2 Set the ambient temperature of the camera to -20℃;

[0036] 1.3 After the camera is powered on, it operates at the specified temperature for 60 minutes to ensure that the internal operating temperature of the camera and detector is consistent with the ambient temperature.

[0037] 1.4 When the ambient temperature is stable, the dark current of the camera is positively correlated with the exposure time. Let t be the minimum exposure time that the camera can be set to. darkMin The maximum exposure time is t brightMax The maximum exposure time is t. darkMax At that time, none of the camera channels experienced saturation. At t darkMin With t darkMax At least m (m≥10) exposure time points t are selected evenly between them. darkExpAt each exposure time point, n (n≥10) dark field images were acquired, and the mean image at each exposure time was calculated.

[0038] 1.5 Calculate the dark field mean image at each exposure time point under the current ambient temperature. Based on the detector's color channel array, the dark field mean image is divided into smaller images for the R, G, and B channels.

[0039] 1.6 Based on the exposure time point t darkExp Dark is the average grayscale value of the pixel in the i-th row and j-th column of the dark field image at each exposure time, representing the dark field mean value of the X and R channels. R By plotting Y and performing a linear fit according to the following formula, the dark field non-uniformity correction coefficient k of the current pixel can be obtained. R_dark With b R_dark By changing the pixel coordinates, you can obtain the dark field non-uniformity correction coefficients for all pixels in the current channel.

[0040] Dark R (i,j)=k R_dark (i,j)*t darkExp +b R_dark (i,j)

[0041] 1.7 Repeat (6), and the same calculation method can be used to obtain the dark field non-uniformity correction coefficients of all pixels in the G and B channels;

[0042] 1.8 Within the temperature range of [-20, 60]℃, change the ambient temperature in 10℃ increments and repeat steps (3)-(7) to calculate the dark field non-uniformity correction coefficient for each channel at each ambient temperature point.

[0043] Step 2: Determining the coefficients of higher-order functions of dark current.

[0044] 2.1 The dark current inside the camera exhibits an exponential relationship with the ambient temperature. When the ambient temperature increases linearly, the dark current inside the camera increases exponentially. Within the temperature range [-20, 60]℃, by changing the ambient temperature in 10℃ increments, the dark current value k of the camera at each ambient temperature can be obtained. T_dark ;

[0045] 2.2 Coefficient k of a single pixel at each temperature point, with ambient temperature T as X. dark Let Y be the value of the dark current. After fitting the higher-order curve according to the following formula, the higher-order function coefficients a and b of the dark current of each pixel can be obtained.

[0046] k dark (i,j)=a*e b*T

[0047] Step 3: Determining the correction coefficient for non-uniformity of radiation response.

[0048] 3.1 Place the camera in a uniformly illuminated environment with the integrating sphere as the sole light source. Set the operating current of the integrating sphere to ensure both the camera and the integrating sphere reach a stable operating state. Turn on the integrating sphere and obtain the minimum exposure time t of the camera under the current integrating sphere illumination conditions. brightMin With maximum exposure time t brightMax Minimum exposure time t brightMin Generally, this refers to the minimum exposure time that the camera can set, and the maximum exposure time t. brightMax This is the maximum exposure time at which all channels of the color camera are unsaturated.

[0049] 3.2 At the minimum exposure time t brightMin With maximum exposure time t brightMax An average of m (m≥10) exposure time points were set between the points, and n (n≥10) bright and dark field images were collected at each exposure time point, and the corresponding mean image was calculated.

[0050] 3.3 Subtract the mean image of the bright field from the mean image of the dark field to obtain the difference image at a single exposure time point;

[0051] 3.4 By linearly fitting the exposure time as X and the gray value of each pixel in the difference image at the corresponding exposure time as Y, the non-uniformity correction coefficients k(i,j) and b(i,j) of the radiometric response of each pixel in the image can be obtained.

[0052] 3.5 Based on the color channel distribution in the detector, the radiation response non-uniformity correction coefficients k, b and kAve, bAve of each channel can be extracted.

[0053] Step 4: Perform joint non-uniformity correction based on environmental conditions, camera exposure time, and various correction coefficients.

[0054] 4.1 Acquire a raw output Bayer image from the camera at a certain temperature T0 and exposure time t0; Substitute the current temperature T0 and exposure time t0 into the formulas in sections 1.6 and 2.2 to calculate the current theoretical dark field Bayer image;

[0055] 4.2 Subtracting the original Bayer image from the theoretical dark-field Bayer image yields the difference image. diff After processing the detector array into channels, the non-uniformity of the radiation response of each pixel in each channel is corrected according to the following formula, resulting in the image after radiation response non-uniformity correction.

[0056]

[0057] 4.3 Image after radiation response non-uniformity correction diffModify Adding this to the background bias yields the final Bayer image after joint correction for dark field and radiometric response inhomogeneity.

[0058] image Modify =image diffModify +b darkAve

[0059] The beneficial effect of this method is that, by employing a correction algorithm that first performs single-sided dark-field correction on the Bayer image based on ambient temperature and exposure time, then performs bilateral radiative response non-uniformity correction within each channel, and finally performs inverse linear restoration of the dark-field signal, it can be widely applied to the non-uniformity correction of color camera output images under different ambient temperatures, exposure times, and lighting conditions. This method provides a new approach to the non-uniformity correction of each channel of a color camera, and also provides a theoretical basis for more accurate color correction of cameras in a wide-temperature environment.

[0060] The above-described joint correction steps for dark field and radiometric response non-uniformity of a color area array camera are merely operational examples of this invention and do not limit the specific content of this invention. Any modifications made without departing from the technical essence of this solution are within the protection scope of this solution.

Claims

1. A method for joint correction of dark field and radiometric response nonuniformity in a color area array camera, characterized in that, Includes the following steps: (1) Place the camera in a dark room, close the protective cover in front of the camera lens, power on the camera, and put the camera in a dark environment. (2) Set the ambient temperature of the camera to T and stabilize for 60 minutes; (3) The minimum exposure time that the camera can be set to is t darkMin The maximum exposure time is t darkMax ; in t darkMin With t darkMax At least m exposure time points t are selected evenly between them. darkExp And acquire n dark field images at each exposure time point; extract dark field images of each color channel R, G, B at all exposure time points according to the distribution of color channels in the detector; (4) In color channel R, at exposure time point t darkExp Let X be the average grayscale value of the pixel in the i-th row and j-th column of the dark field mean image at each exposure time, expressed as Dark. R By plotting Y and performing a linear fit according to the following formula, the dark field non-uniformity correction coefficient k of the current pixel can be obtained. R_dark With b R_dark ; Dark R (i,j)=k R_dark (i,j)*t darkExp +b R_dark (i,j) (5) Change the pixel coordinates to obtain the dark field non-uniformity correction coefficients for all pixels in the current color channel. (6) Repeat steps (4)-(5) to obtain the dark field non-uniformity correction coefficients of all pixels in color channel G and color channel B. (7) Change the ambient temperature and repeat steps (2)-(6) to calculate the dark field non-uniformity correction coefficients for each color channel under all ambient temperatures. (8) Taking temperature T as X, the coefficient k of a single pixel at each temperature point dark Let Y be the dark current coefficients a and b. After performing high-order curve fitting according to the following formula, e is the natural constant. k dark (i,j)=a*e b*T (9) Turn on the integrating sphere, set the operating current, and bring the camera and integrating sphere to a stable operating state; in a uniformly illuminated environment with the integrating sphere as the sole light source, obtain the camera's minimum exposure time t. brightMin With maximum exposure time t brightMax ; (10) At the minimum exposure time t brightMin With maximum exposure time t brightMax An average of m exposure time points are set between each exposure time point. At each exposure time point, n bright and dark field images are collected, and the corresponding mean image is calculated. (11) Subtract the mean image of the bright field from the mean image of the dark field to obtain the difference image at a single exposure time point; (12) Using the exposure time as X and the gray value of each pixel in the difference image at different exposure time points as Y, linear fitting is performed to obtain the radiation response non-uniformity correction coefficients k(i,j) and b(i,j) for each pixel in the image. (13) According to the color channel distribution in the detector, extract the radiation response non-uniformity correction coefficients k, b and kAve, bAve for each channel; (14) Obtain a raw output Bayer image from the camera at a certain temperature environment T0 and exposure time t0; Substitute the current temperature environment T0 and exposure time t0 into the formulas in steps (4) and (8) to calculate the current theoretical dark field Bayer image. (15) Subtract the original output Bayer image from the theoretical dark field Bayer image to obtain the difference image. diff After processing the detector array into channels, the non-uniformity of the radiation response of each pixel in each channel is corrected according to the following formula to obtain the image after non-uniformity correction of radiation response. (16) Add the image after radiation response non-uniformity correction to the background bias of the corresponding pixel to obtain the final Bayer image after joint correction of dark field and radiation response non-uniformity. image Modify =image diffModify +b darkAve 。 2. The method according to claim 1, characterized in that: The ambient temperature T in step (1) refers to the temperature range within which the camera can operate normally.

3. The method according to claim 1, characterized in that: The number of exposure time points and the number of images in steps (3) and (10) must satisfy the following conditions: m≥20, n≥10.

4. The method according to claim 1, characterized in that: In steps (4) and (15), (i,j) represents the coordinates of a single pixel in the image.

5. The method according to claim 1, characterized in that: The minimum exposure time t in step (3) darkMin With maximum exposure time t darkMax These are the minimum and maximum exposure time limits that the camera can set.

6. The correction method according to claim 1, characterized in that: The minimum exposure time t in step (9) brightMin The minimum exposure time that the camera can be set to; the maximum exposure time t brightMax This is the maximum exposure time under the current integrating sphere lighting conditions where none of the channels are exposed.

Images