A stray light correction method for optical remote sensing camera

Abstract

The present invention discloses a stray light correction method for an optical remote sensing camera, comprising: 1. building a test system, illuminating the image plane of the camera in different regions, and acquiring images of each region using a focal plane detector array; 2. image preprocessing; 3. calculating the stray light influence factor of each image illumination region on each pixel in other non-illuminated regions, setting the stray light influence factor of the imaging region to zero, and forming a stray light distribution matrix of each region image by the stray light influence factors of all pixels; 4. the camera captures an image in a real scene, the image partition is consistent with step one, the stray light influence of each region is calculated, and an iterative optimization correction algorithm is used to deduct the sum of the stray light influences of each region from the original measured image, thereby completing the stray light correction. The present invention can realize the stray light correction of an optical remote sensing camera, the method is simple and reliable, has strong versatility and portability, and provides a guarantee for accurate remote sensing data application.

Description

Technical Field

[0001] The invention belongs to the technical field of aerospace remote sensing, and in particular relates to a method for correcting stray light of an optical remote sensing camera. Background Art

[0002] Stray light is light that reaches the detector along the non-imaging light path, thus forming stray light radiation energy and reducing the contrast and clarity of the image. In some serious cases, the target image is even annihilated in the stray light radiation, seriously affecting the quantitative application of remote sensing data. Most optical remote sensing cameras at home and abroad are easily affected by stray light and are interfered by stray light from direct sunlight, which limits the level of quantitative application. Therefore, in order to achieve high-precision remote sensing data applications, it is urgent to study how to suppress and eliminate stray light.

[0003] The main means of stray light suppression are: using matte coating on the surface of the optical-mechanical structure, designing internal and external shading hoods and light-blocking rings, and coating the optical surface with anti-reflection film. After adopting the above suppression methods, stray light measurement is performed. If the stray light is still at a high level, the remaining stray light radiation needs to be corrected. There are two main correction methods: image restoration method and matrix method. The image restoration method is based on the deconvolution algorithm. The actual image is deconvolved with the PSF of the optical system to obtain the image after stray light correction. However, the PSF of the optical system is difficult to measure accurately, and the PSF of different fields of view may not be completely consistent, which brings a huge amount of calculation to the image restoration method. The matrix method was first proposed by Zong and applied to the spectral stray light correction of the spectrometer. Spectral stray light is defined as the response of the remaining wavelength positions outside the central pixel bandwidth, which is similar to the response of the target pixel of the optical remote sensing camera to the remaining pixel positions. Therefore, this method can be applied to the stray light correction of remote sensing cameras. The core is to estimate the stray light response. First, a stray light distribution matrix is ​​constructed. The product of the image response of real scene imaging and the corresponding stray light distribution matrix is ​​considered to be the stray light response. The image of real scene imaging minus the stray light response is the stray light corrected image. However, the stray light distribution matrix is ​​measured in a dark room environment in the laboratory. Under real scene imaging conditions, the stray light response obtained by multiplying the stray light distribution matrix tested in the laboratory by the image response of real scene imaging will overestimate the real signal in the imaging area, resulting in over-correction of stray light. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention proposes a stray light correction method for an optical remote sensing camera, in order to achieve simple and reliable stray light correction for an optical remote sensing camera, improve the versatility and portability of the correction method, and thus provide a guarantee for accurate remote sensing data application.

[0005] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme:

[0006] The present invention provides a method for correcting stray light of an optical remote sensing camera, which comprises the following steps:

[0007] Step 1: Build a stray light test system;

[0008] The stray light testing system is composed of an integrating sphere light source, a field aperture, an optical remote sensing camera, a two-dimensional turntable and a computer. The optical remote sensing camera includes a lens and a photoelectric detector and is in a darkroom environment. The optical remote sensing camera is installed on the two-dimensional turntable.

[0009] The integrating sphere light source, the field aperture, and the optical remote sensing camera are sequentially arranged in the darkroom environment, and the centers of the integrating sphere light source and the field aperture are coaxial with the optical axis of the optical remote sensing camera;

[0010] Step 2: The light emitted by the integrating sphere light source passes through the field aperture and the lens in sequence and reaches the photodetector; the imaging target surface of the photodetector is divided into block illumination area; determine the field aperture size corresponding to each illumination area according to ray tracing; set the short integration time for each illumination area , so that the signal value of each illumination area is non-saturated, so that multiple non-saturated images of each illumination area are obtained by taking pictures of the optical remote sensing camera; and then a long integration time is set for each illumination area , so that the signal value of each illumination area is saturated, so that multiple saturated images of each illumination area are obtained by taking pictures with the optical remote sensing camera;

[0011] Step 3, the computer collects and pre-processes the image acquired by the optical remote sensing camera;

[0012] Step 3.1: After averaging multiple unsaturated images of each illumination area, background subtraction correction is performed, and then nonlinear and non-uniformity correction is performed to obtain a pre-processed unsaturated image of each illumination area, thereby obtaining A preprocessed non-saturated image;

[0013] Step 3.2: After averaging multiple saturated images of each illumination area, background subtraction correction is performed, and then nonlinear and non-uniformity correction is performed to obtain a pre-processed saturated image of each illumination area, thereby obtaining A saturated image after preprocessing;

[0014] Step 4: Set the imaging target surface of the photodetector to have OK Column pixels, let In the preprocessed non-saturated image of the illumination area Line The response value of the column pixel is defined as , then using formula (1) we can get In the preprocessed saturated image of the illumination area Line The true response value of the column pixel :

[0015] ; (1)

[0016] Step 5: Determine the stray light impact factor of each pixel to construct a stray light distribution matrix;

[0017] Step 5.1: Set The response value of the illumination area is defined as , then using formula (2) we can get The stray light impact factor of the illuminated area on the pixel in the i-th row and j-th column in other non-illuminated areas :

[0018] (2)

[0019] Step 5.2: The stray light influence factors of all pixels on the imaging target surface of the photodetector (4) are arranged and composed stray light distribution matrix, where The stray light distribution matrix of the illumination area is recorded as: ;

[0020] Step 6: Stray light correction;

[0021] Step 6.1: Use the optical remote sensing camera to capture images in a real scene, and divide the real scene images into area; suppose the measured response of the real scene image is , and the real response of the real scene image The cumulative effect of stray light composition;

[0022] Step 6.2, define and initialize the number of iterations k=1;

[0023] Step 6.3: Transform the measured response of the real scene image As the true response, and using formula (3) to get the kth stray light estimation response :

[0024] (3)

[0025] Step 6.4: Use equation (4) to obtain the image response after the k-th stray light correction: :

[0026] (4)

[0027] Step 6.5: Use equation (5) to obtain the k+1 stray light estimation response: :

[0028] (5)

[0029] Step 6.6, judgment Is it true? If so, output the image after stray light correction; otherwise, assign k+1 to k and return to step 6.4 to execute sequentially; where, represents the threshold value, and .

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] The present invention uses a combination of image processing, iterative optimization algorithms and other technical means to solve the problems of complex image data processing of image deconvolution algorithm restoration and over-correction of stray light correction by matrix method. It has strong versatility and practicality, can reduce the technical requirements for test instruments and equipment in the process of stray light correction, and ensure test accuracy and save costs. The method of the present invention is simple and reliable, and can meet the needs of stray light correction for various remote sensing cameras. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 Schematic diagram of the stray light testing system of the present invention;

[0033] Figure 2 is a flow chart of the stray light correction method of the present invention;

[0034] Figure 3a This is the 1m aperture integrating sphere image collected in Example 1;

[0035] Figure 3b is the image after stray light correction in Example 1;

[0036] Figure 4a This is a curve diagram of DN value variation before and after stray light correction in the 500th row imaging area in Example 1;

[0037] Figure 4b This is a curve diagram of DN value variation before and after stray light correction in the non-imaging area of ​​the 500th row in Example 1;

[0038] Figure 5a This is the 1.5m integrating sphere image collected in Example 2;

[0039] Figure 5bis the image after stray light correction in Example 2;

[0040] Figure 6a This is a curve diagram of DN value variation before and after stray light correction in the 250th imaging area in Example 2;

[0041] Figure 6b This is a curve diagram of the DN value change before and after the stray light correction in the 250th column of the non-imaging area in Example 2. DETAILED DESCRIPTION

[0042] Example 1

[0043] Stray light correction is performed on an optical remote sensing camera carried by a certain type of satellite with a focal length of 5.54mm, a field of view of 119°, and a working band range of 443-910nm.

[0044] The specific steps of this embodiment are as follows: Figure 2 As shown:

[0045] Step 1: Build a stray light test system, such as Figure 1 As shown:

[0046] The stray light test system is composed of an integrating sphere light source 1, a field aperture 2, an optical remote sensing camera, a two-dimensional turntable 5 and a computer 6. The optical remote sensing camera includes a lens 3 and a photodetector 4 and is in a darkroom environment; the optical remote sensing camera is installed on the two-dimensional turntable 5;

[0047] In a darkroom environment, an integrating sphere light source 1, a field aperture 2, and an optical remote sensing camera are arranged in sequence, and the centers of the integrating sphere light source 1 and the field aperture 2 are coaxial with the optical axis of the optical remote sensing camera.

[0048] Step 2: The light emitted by the integrating sphere light source 1 passes through the field aperture 2 and the lens 3 in sequence and reaches the photodetector 4. After comprehensively evaluating the stray light and noise levels, the imaging target surface is divided into According to the ray tracing, determine the field aperture size corresponding to each illumination area; set the short integration time of each illumination area to 30ms, so that the signal value of each illumination area is non-saturated, so that 50 non-saturated images of each illumination area are obtained by the optical remote sensing camera; then set the long integration time of each illumination area to 1500ms, so that the signal value of each illumination area is saturated, so that 50 saturated images of each illumination area are obtained by the optical remote sensing camera;

[0049] Step 3, the computer 5 collects and pre-processes the images acquired by the optical remote sensing camera;

[0050] Step 3.1: The 50 unsaturated images of each illumination area are averaged, and background subtraction correction is performed, and then nonlinear and non-uniformity correction is performed to obtain a pre-processed unsaturated image of each illumination area, thereby obtaining A preprocessed non-saturated image;

[0051] Step 3.2: The 50 saturated images of each illumination area are averaged, and background subtraction correction is performed, and then nonlinear and non-uniformity correction is performed to obtain a pre-processed saturated image of each illumination area, thereby obtaining A saturated image after preprocessing;

[0052] Step 4: Set the imaging target surface of the photodetector 4 to have OK Column pixels, let In the preprocessed non-saturated image of the illumination area Line The response value of the column pixel is defined as , then using formula (1) we can get In the preprocessed saturated image of the illumination area Line The true response value of the column pixel :

[0053] ; (1)

[0054] Step 5: Determine the stray light impact factor of each pixel to construct a stray light distribution matrix;

[0055] Step 5.1: Set The response value of the illumination area is defined as , then using formula (2) we can get The stray light impact factor of the illuminated area on the pixel in the i-th row and j-th column in other non-illuminated areas :

[0056] (2)

[0057] Step 5.2: The stray light influence factors of all pixels on the imaging target surface of the photodetector (4) are arranged and composed of stray light distribution matrix, where The stray light distribution matrix of the illumination area is recorded as: ;

[0058] Step 6: Stray light correction:

[0059] Step 6.1: Use an optical remote sensing camera to take an image of an integrating sphere with an aperture of 1 m, such as Figure 3a As shown, the integrating sphere image is divided into area. Assume that the measurement response of the integrating sphere image is , and the real response of the integrating sphere image The cumulative effect of stray light composition;

[0060] Step 6.2, define and initialize the number of iterations k=1;

[0061] Step 6.3: Transform the measurement response of the integrating sphere image As the true response, and using formula (3) to get the kth stray light estimation response :

[0062] (3)

[0063] Step 6.4: Use equation (4) to obtain the image response after the k-th stray light correction: :

[0064] (4)

[0065] Step 6.5: Use equation (5) to obtain the k+1 stray light estimation response: :

[0066] (5)

[0067] Step 6.6, judgment Is it true? If so, output the image after stray light correction; otherwise, assign k+1 to k and return to step 6.4 to execute sequentially; where, represents the threshold value, and

[0068] This embodiment runs a total of 100 iteration cycles and outputs a stray light corrected image, such as Figure 3b In order to compare the changes of the image before and after correction, Figure 4a and Figure 4b The DN value change curves of the imaging area of ​​the 500th line of the image before and after the stray light correction and the DN value change curves of the non-imaging area before and after the stray light correction are respectively given. It can be seen from the figure that this embodiment corrects 96% of the stray light effect on average.

[0069] Example 2

[0070] Stray light correction is performed on an optical remote sensing camera carried by a certain type of satellite with a focal length of 4.8mm, a field of view of 107°, and a working band range of 490-910nm.

[0071] The specific steps of this embodiment are as follows:

[0072] Step 1: Building a stray light testing system, which is consistent with step 1 in Example 1.

[0073] Step 2: In this embodiment, the imaging target surface is divided into Block illumination area. According to ray tracing, determine the field aperture size corresponding to each illumination area. Set the short integration time of each illumination area to 20ms, so that the signal value of each illumination area is non-saturated, so that 100 non-saturated images of each illumination area are obtained by the optical remote sensing camera; then set the long integration time of each illumination area to 1500ms, so that the signal value of each illumination area is saturated, so that 100 saturated images of each illumination area are obtained by the optical remote sensing camera;

[0074] Step 3: The computer collects and pre-processes the images acquired by the optical remote sensing camera;

[0075] Step 3.1: The 100 unsaturated images of each illumination area are averaged, and then background subtraction correction is performed, and then nonlinear and non-uniformity correction is performed to obtain a pre-processed unsaturated image of each illumination area, thereby obtaining A preprocessed non-saturated image;

[0076] Step 3.2: The 100 saturated images of each illumination area are averaged and then subjected to background subtraction correction and then nonlinear and non-uniformity correction to obtain a pre-processed saturated image of each illumination area, thereby obtaining A saturated image after preprocessing;

[0077] Step 4: Set the imaging target surface of the photodetector to OK Column pixels, let In the preprocessed non-saturated image of the illumination area Line The response value of the column pixel is defined as , then using formula (1) we can get In the preprocessed saturated image of the illumination area Line The true response value of the column pixel :

[0078] ; (1)

[0079] Step 5: Determine the stray light impact factor of each pixel to construct a stray light distribution matrix;

[0080] Step 5.1: Set The response value of the illumination area is defined as , then using formula (2) we can get The stray light impact factor of the illuminated area on the pixel in the i-th row and j-th column in other non-illuminated areas :

[0081] (2)

[0082] Step 5.2: The stray light influence factors of all pixels on the imaging target surface of the photodetector are arranged and composed stray light distribution matrix, where The stray light distribution matrix of the illumination area is recorded as: ;

[0083] Step 6: Stray light correction:

[0084] Step 6.1: Use an optical remote sensing camera to take an image of an integrating sphere with an aperture of 1.5 m, such as Figure 5a As shown, the integrating sphere image is divided into area. Assume that the measurement response of the integrating sphere image is , and the real response of the integrating sphere image The cumulative effect of stray light composition;

[0085] Step 6.2, define and initialize the number of iterations k=1;

[0086] Step 6.3: Transform the measurement response of the integrating sphere image As the true response, and using formula (3) to get the kth stray light estimation response :

[0087] (3)

[0088] Step 6.4: Use equation (4) to obtain the image response after the k-th stray light correction: :

[0089] (4)

[0090] Step 6.5: Use equation (5) to obtain the k+1 stray light estimation response: :

[0091] (5)

[0092] Step 6.6, judgment Is it true? If so, output the image after stray light correction; otherwise, assign k+1 to k and return to step 6.4 to execute sequentially; where, represents the threshold value, and ;

[0093] This embodiment runs a total of 150 iterations and outputs a stray light corrected image, such as Figure 5b In order to compare the changes of the image before and after correction, Figure 6a and Figure 6b The DN value change curves before and after the stray light correction in the imaging area of ​​the 250th column of the image and the DN value change curves before and after the stray light correction in the non-imaging area are respectively given. It can be seen from the figure that this embodiment corrects 92% of the stray light effect on average.

Claims

1. A stray light correction method for an optical remote sensing camera, characterized in that: The steps include: Step 1: Build a stray light test system; The stray light testing system is composed of an integrating sphere light source (1), a field aperture (2), an optical remote sensing camera, a two-dimensional turntable (5), and a computer (6); the optical remote sensing camera comprises a lens (3), a photoelectric detector (4), and is located in a darkroom environment; the optical remote sensing camera is installed on the two-dimensional turntable (5); The integrating sphere light source (1), the field diaphragm (2), and the optical remote sensing camera are arranged in sequence in the darkroom environment, and the centers of the integrating sphere light source (1) and the field diaphragm (2) are coaxial with the optical axis of the optical remote sensing camera; Step 2: The light emitted by the integrating sphere light source (1) passes through the field aperture (2) and the lens (3) in sequence and reaches the photodetector (4); the imaging target surface of the photodetector (4) is divided into Block lighting area; According to the ray tracing, determine the field aperture size corresponding to each illumination area; set the short integration time for each illumination area , so that the signal value of each illumination area is unsaturated, so that a plurality of unsaturated images of each illumination area are obtained by taking pictures with the optical remote sensing camera; Then set the long integration time for each lighting area , so that the signal value of each illumination area is saturated, so that multiple saturated images of each illumination area are obtained by taking pictures with the optical remote sensing camera; Step 3, the computer (6) collects and pre-processes the image acquired by the optical remote sensing camera; Step 3.1: After averaging multiple unsaturated images of each illumination area, background subtraction correction is performed, and then nonlinear and non-uniformity correction is performed to obtain a pre-processed unsaturated image of each illumination area, thereby obtaining A preprocessed non-saturated image; Step 3.2: After averaging multiple saturated images of each illumination area, background subtraction correction is performed, and then nonlinear and non-uniformity correction is performed to obtain a pre-processed saturated image of each illumination area, thereby obtaining A saturated image after preprocessing; Step 4: Setting the imaging target surface of the photodetector (4) to have OK Column pixels, let In the preprocessed non-saturated image of the illumination area Line The response value of the column pixel is defined as , then using formula (1) we can get In the preprocessed saturated image of the illumination area Line The true response value of the column pixel : ； (1) Step 5: Determine the stray light impact factor of each pixel to construct a stray light distribution matrix; Step 5.1: Set The response value of the illumination area is defined as , then using formula (2) we can get The stray light impact factor of the illuminated area on the pixel in the i-th row and j-th column in other non-illuminated areas : (2) Step 5.2: The stray light influence factors of all pixels on the imaging target surface of the photodetector (4) are arranged and composed stray light distribution matrix, where The stray light distribution matrix of the illumination area is recorded as: ； Step 6: Stray light correction; Step 6.1: Use the optical remote sensing camera to capture images in a real scene, and divide the real scene images into area; suppose the measured response of the real scene image is , and the real response of the real scene image The cumulative effect of stray light composition; Step 6.2, define and initialize the number of iterations k=1; Step 6.3: Transform the measured response of the real scene image As the true response, and using formula (3) to get the kth stray light estimation response : (3) Step 6.4: Use equation (4) to obtain the image response after the k-th stray light correction: : (4) Step 6.5: Use equation (5) to obtain the k+1 stray light estimation response: : (5) Step 6.6, judgment Is it true? If so, output the image after stray light correction; otherwise, assign k+1 to k and return to step 6.4 to execute sequentially; where, represents the threshold value, and .

Images