Dark field image correction method and device
By generating temperature and exposure time correction parameters, and correcting based on the fitting curve of the dark field image, the uniformity and noise problems of dark field images at different temperatures and exposure times are solved, and high-quality imaging within the entire interval of Sensor is achieved.
Patent Information
- Application Number
- CN202411468885.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-21
- Publication Date
- 2025-01-14
AI Technical Summary
The prior art cannot effectively correct dark field images at different temperatures and exposure times, resulting in poor image uniformity, presence of stripe noise, poor edge and corner brightness correction effect, and cannot guarantee the DSNU index within the entire Sensor interval.
By generating a fitting curve with exposure time based on the reference temperature and image cell grayscale value at different temperatures, and obtaining temperature and exposure time correction parameters, the full interval background correction of dark field images is achieved, and the bright light at the edges and corners are eliminated and noise is reduced. Influence.
Improves the uniformity of dark field images, weakens fringe noise, ensures DSNU indicators at each temperature, improves imaging quality, avoids over-correction, and ensures the imaging performance of Sensor in the entire interval.
Smart Images
Figure CN119316740A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of image processing, and in particular to a dark field image correction method and device. Background Art
[0002] Dark field images refer to images collected without incident light or when the camera is covered with a dust cover. Due to the influence of temperature and exposure time, dark field images at different temperatures and exposure times have different degrees of noise, including dark current, streak noise, corner light, etc. The image uniformity is poor, which seriously affects the imaging quality.
[0003] Due to the inconsistency of chip characteristics and the photosensitivity characteristics of each pixel of the sensor, the dark current at different temperatures varies greatly. The lower the temperature, the smaller the dark current. In addition, dark field images under long exposure at the same temperature are prone to bright light at the corners, which deteriorates the imaging effect and causes image signal loss in low-light scenes. Therefore, the non-uniformity correction effect of the dark field image directly affects the imaging quality of the sensor.
[0004] For the non-uniformity correction of dark field images, the commonly used technical means is to collect dark field images with different exposure times at a certain temperature, linearly fit the grayscale values of each pixel at different exposure times, obtain the correction parameters (slope and intercept) of each pixel, and then correct the dark field image according to the correction parameters. This method can only correct dark field images at a specific temperature. If the operating temperature of the camera changes, the correction effect will be greatly reduced, the stripe noise cannot be removed, and the DSNU (dark signal non-uniformity) index will deteriorate. In addition, the dark current of the image after correction is prone to over-correction, which changes the original characteristics of the chip, and the corner light correction effect is poor, resulting in data loss.
[0005] The above commonly used technical means have the following disadvantages: 1: Dark field non-uniformity correction can only be achieved at a specific temperature. The correction effect is greatly reduced at other temperatures, with obvious stripe noise and poor dark field background uniformity; 2: After correction, the dark current of the sensor is prone to over-correction, and the correction effect of corner light is poor; 3: The DSNU indicator cannot be guaranteed in all operating temperature ranges of the Sensor. Summary of the invention
[0006] The technical problem to be solved by the present invention is to provide a dark field image correction method and device.
[0007] In order to solve the above technical problems, a technical solution adopted by the present invention is: A dark field image correction method comprises the following steps: Generate a dark field correction parameter according to a temperature correction parameter, an exposure time correction parameter and a correction temperature; wherein the exposure time correction parameter is obtained based on a first fitting curve, a second fitting curve and a correction temperature, the first fitting curve is a fitting curve of changes in image pixel grayscale values with exposure time at a reference temperature, the second fitting curve is a fitting curve of changes in image pixel grayscale values with exposure time at different temperatures, and the correction temperature is a temperature when a dark field image to be corrected is generated; The dark field image to be corrected is corrected according to the dark field correction parameters.
[0008] Another technical solution adopted by the present invention is: A dark field image correction device comprises a correction parameter generation module and a correction module; The correction parameter generation module is used to obtain temperature correction parameters and exposure time correction parameters; The correction module is used to correct the dark field image to be corrected according to the temperature correction parameter and the exposure time correction parameter; The correction parameter generation module and the correction module jointly implement the steps in the dark field image correction method described above.
[0009] The beneficial effects of the present invention are: The present invention provides a dark field image correction method and device, which can generate dark field correction parameters based on the fitting curve of the gray value of the image pixel at the reference temperature and the exposure time, and the fitting curve of the gray value of the image pixel at different temperatures and the exposure time, so as to achieve the full range background correction of the sensor working temperature and exposure time, improve the uniformity of the dark field image, reduce the stripe noise, and ensure the DSNU index at various temperatures. In addition, the background dynamic correction based on the combination of temperature and exposure time can largely eliminate the corner light of the image, weaken the influence of noise, and achieve higher imaging quality. BRIEF DESCRIPTION OF THE DRAWINGS
[0010] Figure 1 A flowchart of a dark field image correction method of the present invention; Figure 2 It is a structural block diagram of a dark field image correction device of the present invention; Figure 3 A flowchart of the steps of a dark field image correction method of the present invention; Description of labels: 1. A dark field image correction device; 2. A correction parameter generation module; 3. A correction module. DETAILED DESCRIPTION
[0011] In order to explain the technical content, achieved objectives and effects of the present invention in detail, the following is an explanation in combination with the implementation modes and the accompanying drawings.
[0012] Please refer to Figure 1 and Figure 3 The present invention provides a dark field image correction method, comprising the following steps: Generate a dark field correction parameter according to a temperature correction parameter, an exposure time correction parameter and a correction temperature; wherein the exposure time correction parameter is obtained based on a first fitting curve, a second fitting curve and a correction temperature, the first fitting curve is a fitting curve of changes in image pixel grayscale values with exposure time at a reference temperature, the second fitting curve is a fitting curve of changes in image pixel grayscale values with exposure time at different temperatures, and the correction temperature is a temperature when a dark field image to be corrected is generated; The dark field image to be corrected is corrected according to the dark field correction parameters.
[0013] From the above description, it can be seen that the beneficial effects of the present invention are: The dark field image correction method provided by the present invention can generate dark field correction parameters based on the fitting curve of the gray value of the image pixel at the reference temperature and the exposure time, and the fitting curve of the gray value of the image pixel at different temperatures and the exposure time, so as to realize the full range background correction of the sensor working temperature and exposure time, improve the uniformity of the dark field image, reduce the stripe noise, and ensure the DSNU index at various temperatures. In addition, the background dynamic correction based on the combination of temperature and exposure time can largely eliminate the corner light of the image, weaken the influence of noise, and have higher imaging quality.
[0014] Further, the exposure time correction parameter is obtained based on the first fitting curve, the second fitting curve and the correction temperature, including: generating the first fitting curve, and using the reference correction slope of the image pixel corresponding to the first fitting curve as the reference correction parameter; Generate the second fitting curve, obtain a plurality of correction slope averages corresponding to different temperatures, and obtain an exponential curve equation of the change of the correction slope average with temperature by fitting; An exposure time correction parameter is generated according to the reference correction parameter, the exponential curve equation and the correction temperature.
[0015] Furthermore, the first fitting curve is generated by dark field images sampled at different exposure times at a reference temperature, and the second fitting curve includes a plurality of third fitting curves corresponding to changes in image pixel grayscale values at different temperatures with exposure time, and each of the third fitting curves is generated by sampled dark field images at the same temperature and different exposure times; and / or The exponential curve equation is obtained by fitting the curve of the correction slope mean k_et_mean changing with temperature according to the second fitting curve: ; Wherein, a1, a2, a3, a4 are exponential equation parameters, T is the temperature variable; and / or The exposure time correction parameters are: ; Wherein, k_et_std represents the reference correction slope, k_et represents the exposure time correction parameter, row represents the image row coordinate, col represents the image column coordinate, and T represents the correction temperature.
[0016] It can be seen from the above description that based on the above method, the exposure time correction parameters are specifically generated.
[0017] Furthermore, the temperature correction parameters are obtained based on a temperature fitting curve of image pixel grayscale values changing with temperature.
[0018] Furthermore, the temperature fitting curve is generated by sampling dark field images at the same exposure time and different temperatures; and / or The temperature correction parameters include the correction slope and intercept of the temperature fitting curve corresponding to the image pixel.
[0019] From the above description, it can be known that the temperature correction parameters are obtained based on the temperature fitting curve of the gray value of the image pixel changing with temperature. Specifically, the correction slope and intercept of the temperature fitting curve corresponding to the image pixel are used as the temperature correction parameters.
[0020] Furthermore, the generation of dark field correction parameters includes: Generate a dark field correction parameter according to the temperature correction parameter, the exposure time correction parameter, the correction temperature and the correction exposure time; The correction exposure time is the exposure time when the dark field image to be corrected is generated.
[0021] It can be seen from the above description that the dark field correction parameter is generated according to the temperature correction parameter, the exposure time correction parameter, and the correction temperature and the correction exposure time.
[0022] Furthermore, after generating the dark field correction parameters and before correcting the dark field image to be corrected according to the dark field correction parameters, the method further includes: The temperature correction parameter and the exposure time correction parameter are controlled to exceed the limit by means of threshold limits.
[0023] Furthermore, the temperature correction parameter and the exposure time correction parameter are controlled by threshold limiting, including: Calculating the mean of the slope parameter in the temperature correction parameter and the mean of the exposure time correction parameter corresponding to the correction temperature to obtain a slope mean and an exposure time correction mean; According to the slope mean and the preset bias, the threshold value of the slope parameter is determined, and the slope parameter is controlled for over-limit; according to the exposure time correction mean and the preset bias proportional coefficient, the threshold value of the exposure time correction parameter is determined, and the exposure time correction parameter is controlled for over-limit.
[0024] From the above description, it can be seen that before correcting the dark field image to be corrected according to the dark field correction parameters, the temperature correction parameters and the exposure time correction parameters are first controlled to prevent over-correction due to excessive correction slopes of some pixels.
[0025] Furthermore, when correcting the dark field image to be corrected, it includes: The compensation process is performed according to a preset exposure compensation coefficient, where the exposure compensation coefficient has a value range of 0 to 16.
[0026] From the above description, it can be seen that when the dark field image to be corrected is corrected, the preset exposure compensation coefficient is also used for compensation processing. The exposure compensation coefficient can basically eliminate the bright light at the corners of the dark field background after correction, greatly reduce the stripe noise, make the dark current curve more linear, and avoid over-correction.
[0027] Please refer to Figure 2 , a dark field image correction device, comprising a correction parameter generation module and a correction module; The correction parameter generation module is used to obtain temperature correction parameters and exposure time correction parameters; The correction module is used to correct the dark field image to be corrected according to the temperature correction parameter and the exposure time correction parameter; The correction parameter generation module and the correction module jointly implement the steps in the dark field image correction method described above.
[0028] The dark field image correction device provided by the present invention can generate dark field correction parameters based on the fitting curve of the gray value of the image pixel at the reference temperature and the exposure time, and the fitting curve of the gray value of the image pixel at different temperatures and the exposure time, so as to realize the full range background correction of the sensor working temperature and exposure time, improve the uniformity of the dark field image, reduce the stripe noise, and ensure the DSNU index at various temperatures. In addition, the background dynamic correction based on the combination of temperature and exposure time can largely eliminate the bright light at the corners of the image, weaken the influence of noise, and achieve higher imaging quality.
[0029] A dark field image correction method and device of the present invention are suitable for non-uniformity correction of dark field images to improve the imaging quality of a camera.
[0030] Please refer to Figure 1 and Figure 3 , Embodiment 1 of the present invention is: A dark field image correction method comprises the following steps: A dark field correction parameter is generated according to a temperature correction parameter, an exposure time correction parameter and a correction temperature; wherein the exposure time correction parameter is obtained based on a first fitting curve, a second fitting curve and a correction temperature, the first fitting curve is a fitting curve of changes in image pixel grayscale values with exposure time at a reference temperature, the second fitting curve is a fitting curve of changes in image pixel grayscale values with exposure time at different temperatures, and the correction temperature is a temperature when a dark field image to be corrected is generated.
[0031] In this embodiment, the camera is placed in a dark field environment (for example, the camera can be covered with a dust cover), different temperatures and exposure times are set, and corresponding dark field images are obtained, where x can be 10 or other values. The temperature can be set to m segments, where m can be 3 or other values, for example, the lowest working temperature, the middle working temperature, and the highest working temperature can be set; the exposure time can be set to n segments, where n can be 2 or other values, for example, dark field images can be taken at the exposure time (short exposure) required for calculating DSNU and the maximum working exposure time (long exposure).
[0032] The temperature correction parameters are obtained based on a temperature fitting curve of image pixel grayscale values varying with temperature, wherein the temperature fitting curve is generated by sampling dark field images at the same exposure time and different temperatures; the temperature correction parameters include a correction slope and an intercept of the temperature fitting curve corresponding to the image pixel.
[0033] In this embodiment, the temperature correction parameters are generated: The mean images of n dark field images at m sampling temperatures with the same exposure time are calculated respectively to obtain m groups of mean images, wherein the exposure time is the exposure time required for calculating DSNU (10 ms in this embodiment). According to the gray value fitting curve of each pixel at m temperatures in the mean image, specifically a linear curve, the correction slope k_temp and intercept b_temp of each pixel changing with temperature are obtained as temperature correction parameters.
[0034] The exposure time correction parameter is obtained based on the first fitting curve, the second fitting curve and the correction temperature, including: The first fitting curve is generated, and the reference correction slope of the image pixel corresponding to the first fitting curve is used as the reference correction parameter; the first fitting curve is generated by dark field images sampled at different exposure times at a reference temperature.
[0035] In this embodiment, n groups of mean images are generated based on dark field images at a certain temperature (reference correction temperature, 0°C in this embodiment) at n exposure times, and a curve of the grayscale value of each pixel changing with the exposure time is fitted, i.e., the first fitting curve. Specifically, it is a linear curve, and the correction slope k_et_std (i.e., the reference correction slope) and the correction slope mean k_et_mean of each pixel changing with the exposure time at this time are saved. The reference correction slope is used as the reference correction parameter.
[0036] Generate the second fitting curve, obtain multiple correction slope means corresponding to different temperatures, and fit to obtain an exponential curve equation of the correction slope mean changing with temperature; the second fitting curve includes multiple third fitting curves corresponding to the grayscale value of the image pixel at different temperatures changing with the exposure time, and each of the third fitting curves is generated by a sampled dark field image at the same temperature and different exposure times.
[0037] In this embodiment, since the correction slope is different at different temperatures, it is necessary to generate a curve of the change of the mean correction slope with temperature. The mean correction slope k_et_mean at different temperatures is solved, and the mean correction slope k_et_mean-temperature curve is fitted, which is specifically an exponential curve. Then the change trend of the correction slope of each pixel with temperature also satisfies the exponential relationship, that is, the exponential curve equation.
[0038] Specifically, the exponential curve equation is obtained by fitting the curve of the correction slope mean k_et_mean changing with temperature according to the second fitting curve: ; Among them, a1, a2, a3, a4 are exponential equation parameters, and T is the temperature variable.
[0039] The exposure time correction parameters are: ; Wherein, k_et_std represents the reference correction slope, k_et represents the exposure time correction parameter, row represents the image row coordinate, col represents the image column coordinate, and T represents the correction temperature.
[0040] A dark field correction parameter is generated according to the temperature correction parameter, the exposure time correction parameter, the correction temperature and the correction exposure time; the correction exposure time is the exposure time when the dark field image to be corrected is generated.
[0041] In this embodiment, the corresponding dark field background, that is, the dark field correction parameter, is generated according to the current temperature and exposure time. The dark field correction parameter changes dynamically with the temperature and exposure time, as follows: ; ; ; Among them, BG_temp represents the dark field background of the Sensor at different temperatures with an exposure time of etDSNU, BG_et represents the background increment of the Sensor with exposure time at different temperatures, et represents the current exposure time, etDSNU represents the exposure time required to calculate DSNU, which is 10ms in this embodiment, and BG represents the dark field background corresponding to the corrected temperature and corrected exposure time, that is, the dark field correction parameter.
[0042] The dark field image to be corrected is corrected according to the dark field correction parameters.
[0043] The original dark field data output by the sensor is corrected according to the generated dark field background BG. In this embodiment, when the dark field image to be corrected is corrected, it includes: The compensation process is performed according to a preset exposure compensation coefficient, where the exposure compensation coefficient has a value range of 0 to 16.
[0044] That is, in order to prevent over-correction and data loss, a bias needs to be accumulated after correction, as follows: ; Among them, src_calib_data represents the calibrated dark field background, offset is the set offset, and comp is the preset exposure compensation coefficient.
[0045] In this embodiment, the exposure compensation coefficient comp ranges from 0 to 16.
[0046] After correction in the above manner, the exposure compensation amount can basically eliminate the bright light in the corners of the dark field background after correction, greatly reduce the stripe noise, make the dark current curve more linear, and there is no over-correction phenomenon.
[0047] Table 1 below shows the DSNU index of the sensor at various temperatures calculated using the method of the present invention. From the data, it can be seen that the difference in DSNU index at various temperatures is small, and the image uniformity is higher.
[0048]
[0049] Table 1 Embodiment 2 of the present invention is: A dark field image correction method is different from the first embodiment in that, after generating dark field correction parameters, before correcting the dark field image to be corrected according to the dark field correction parameters, the method further includes: The temperature correction parameter and the exposure time correction parameter are controlled to exceed the limit by means of threshold limits.
[0050] The temperature correction parameter and the exposure time correction parameter are controlled by threshold limiting, including: Calculating the mean of the slope parameter in the temperature correction parameter and the mean of the exposure time correction parameter corresponding to the correction temperature to obtain a slope mean and an exposure time correction mean; According to the slope mean and the preset bias, the threshold value of the slope parameter is determined, and the slope parameter is controlled for over-limit; according to the exposure time correction mean and the preset bias proportional coefficient, the threshold value of the exposure time correction parameter is determined, and the exposure time correction parameter is controlled for over-limit.
[0051] In this embodiment, the slope parameter in the temperature correction parameter and the mean value of the exposure time correction parameter at the current temperature of the Sensor are calculated respectively to obtain k_tempave and k_etave; then the points in k_temp that exceed k_tempave±tempoffset are set to k_tempave (tempoffset is the temperature correction parameter threshold, which is set to 3 in this embodiment), and the points in k_et that exceed r*k_etave are set to r*k_etave (r is the proportional coefficient, which is set to 10 in this embodiment) to prevent over-correction due to excessive correction slope of some pixels.
[0052] Please refer to Figure 2 , Embodiment 3 of the present invention is: A dark field image correction device 1, comprising a correction parameter generation module 2 and a correction module 3; The correction parameter generation module 2 is used to obtain temperature correction parameters and exposure time correction parameters; The correction module 3 is used to correct the dark field image to be corrected according to the temperature correction parameter and the exposure time correction parameter; The correction parameter generation module 2 and the correction module 3 jointly implement the steps in a dark field image correction method described in any one of the above embodiments 1 or 2.
[0053] In summary, the dark field image correction method and device provided by the present invention can correct the correction parameters of temperature and exposure time according to the dark field image correction under different temperatures and exposure times, and realize the dynamic generation of dark field background based on the current working temperature and exposure time of the sensor, thereby realizing the full-range background correction within the working temperature and exposure time of the sensor, improving the uniformity of the dark field image, reducing stripe noise, and ensuring the DSNU index at various temperatures. In addition, the background dynamic correction based on the combination of temperature and exposure time can eliminate the bright light at the corners of the image to a large extent, weaken the influence of noise, and improve the imaging quality. And by introducing the exposure compensation amount, it is ensured that the corrected dark current curve is more in line with the chip characteristics and has higher linearity. This method is simple to implement and has a high application prospect.
[0054] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent transformations made using the contents of the present invention's specification and drawings, or directly or indirectly applied in related technical fields, are also included in the patent protection scope of the present invention.
Claims
1. A dark field image correction method, characterized in that: The following steps are involved: Generate a dark field correction parameter according to a temperature correction parameter, an exposure time correction parameter and a correction temperature; wherein the exposure time correction parameter is obtained based on a first fitting curve, a second fitting curve and a correction temperature, the first fitting curve is a fitting curve of changes in image pixel grayscale values with exposure time at a reference temperature, the second fitting curve is a fitting curve of changes in image pixel grayscale values with exposure time at different temperatures, and the correction temperature is a temperature when a dark field image to be corrected is generated; The dark field image to be corrected is corrected according to the dark field correction parameters.
2. A dark field image correction method according to claim 1, characterized in that: The exposure time correction parameter is obtained based on the first fitting curve, the second fitting curve and the correction temperature, including: generating the first fitting curve, and using the reference correction slope of the image pixel corresponding to the first fitting curve as the reference correction parameter; Generate the second fitting curve, obtain a plurality of correction slope averages corresponding to different temperatures, and obtain an exponential curve equation of the change of the correction slope average with temperature by fitting; An exposure time correction parameter is generated according to the reference correction parameter, the exponential curve equation and the correction temperature.
3. A dark field image correction method according to claim 2, characterized in that: The first fitting curve is generated by dark field images sampled at different exposure times at a reference temperature, the second fitting curve includes a plurality of third fitting curves corresponding to changes in image pixel grayscale values at different temperatures with exposure time, each of the third fitting curves is generated by sampled dark field images at the same temperature and different exposure times; and / or The exponential curve equation is obtained by fitting the curve of the correction slope mean k_et_mean changing with temperature according to the second fitting curve: ； Wherein, a1, a2, a3, a4 are exponential equation parameters, T is the temperature variable; and / or The exposure time correction parameters are: ； Wherein, k_et_std represents the reference correction slope, k_et represents the exposure time correction parameter, row represents the image row coordinate, col represents the image column coordinate, and T represents the correction temperature.
4. A dark field image correction method according to claim 1, characterized in that: The temperature correction parameters are obtained based on a temperature fitting curve of the grayscale value of an image pixel changing with temperature.
5. A dark field image correction method according to claim 4, characterized in that: The temperature fitting curve is generated by sampling dark field images at the same exposure time and different temperatures; and / or The temperature correction parameters include the correction slope and intercept of the temperature fitting curve corresponding to the image pixel.
6. A dark field image correction method according to claim 1, characterized in that: The generation of dark field correction parameters includes: Generate a dark field correction parameter according to the temperature correction parameter, the exposure time correction parameter, the correction temperature and the correction exposure time; The correction exposure time is the exposure time when the dark field image to be corrected is generated.
7. A dark field image correction method according to claim 6, characterized in that: After generating the dark field correction parameters and before correcting the dark field image to be corrected according to the dark field correction parameters, the method further includes: The temperature correction parameter and the exposure time correction parameter are controlled to exceed the limit by means of threshold limits.
8. A dark field image correction method according to claim 7, characterized in that: The temperature correction parameter and the exposure time correction parameter are controlled by threshold limiting, including: Calculating the mean of the slope parameter in the temperature correction parameter and the mean of the exposure time correction parameter corresponding to the correction temperature to obtain a slope mean and an exposure time correction mean; According to the slope mean and the preset bias, the threshold value of the slope parameter is determined, and the slope parameter is controlled for over-limit; according to the exposure time correction mean and the preset bias proportional coefficient, the threshold value of the exposure time correction parameter is determined, and the exposure time correction parameter is controlled for over-limit.
9. A dark field image correction method according to claim 1, characterized in that: Also includes: When correcting the dark field image to be corrected, the method includes: The compensation process is performed according to a preset exposure compensation coefficient, where the exposure compensation coefficient has a value range of 0 to 16.
10. A dark field image correction device, characterized in that: It includes a correction parameter generation module and a correction module; The correction parameter generation module is used to obtain temperature correction parameters and exposure time correction parameters; The correction module is used to correct the dark field image to be corrected according to the temperature correction parameter and the exposure time correction parameter; The correction parameter generation module and the correction module jointly implement the steps in the dark field image correction method described in any one of claims 1 to 9 above.