A high-precision radiometric calibration method for very high sensitivity infrared sea surface imaging data

Abstract

A high-precision radiometric calibration method for very high sensitivity infrared sea imaging data is proposed. Firstly, the blackbody calibration data on the infrared remote sensor satellite are used to calculate the radiometric calibration coefficient and detect the blind pixels. Then, the image data are corrected by the primary relative radiometric correction and blind pixel compensation correction. Secondly, a radiometric correction method based on small sample statistics is proposed to remove the residual stripe noise after the primary correction, which further improves the radiometric quality of the infrared image. Finally, the blind pixels in the image are detected and removed.

Description

Technical Field

[0001] This invention relates to a high-precision radiometric calibration method for very sensitive infrared imaging data of the sea, belonging to the field of aerospace remote sensing technology. Background Technology

[0002] Marine infrared remote sensing has wide applications in sea surface temperature anomaly monitoring, fisheries guidance, sea surface oil pollution detection, and underwater target detection. With its increasing application, the demand for remote sensors with high detection sensitivity is becoming increasingly strong. Infrared detectors with a temperature sensitivity better than 10 mK are generally called very high sensitivity infrared detectors. At such high temperature sensitivity, non-uniformity accounts for a large proportion of the total noise. The greater the proportion of non-uniformity in the total noise, the more significant the attenuation of sensitivity becomes, even affecting the effective detection and identification of weak targets. Furthermore, because the temperature distribution at the sea surface is more uniform than that on land, the residual stripe noise at the sea surface is more visually apparent under the same residual non-uniformity. Therefore, the non-uniformity requirements for sea surface remote sensing images are higher, both for visual effectiveness and subsequent target detection.

[0003] Currently available radiometric calibration methods for infrared remote sensing data can be mainly divided into two categories: 1) radiometric calibration methods based on radiation sources; and 2) radiometric calibration methods based on scenarios.

[0004] For very high-sensitivity infrared remote sensors, the large optical aperture makes it difficult to achieve full-aperture radiometric calibration of the entire optical path on the satellite. Generally, a small on-board blackbody is set near the primary image plane to perform half-optical-path full-aperture radiometric calibration. However, the on-board calibration blackbody is limited by energy consumption and volume, and can generally only achieve one or two temperature points for single-point or two-point radiometric calibration. At the same time, due to the nonlinear response characteristics of infrared remote sensors, the images corrected based on the on-board calibration blackbody will have residual stripe noise, which cannot meet the radiometric calibration accuracy of very high-sensitivity infrared imaging data of the sea.

[0005] Scene-based radiometric calibration methods require large sample data for statistical analysis and are generally used for radiometric calibration of visible and near-infrared remote sensing data. Since the non-uniform noise of an infrared remote sensor changes with each power-on, the only statistically available data sample for infrared remote sensing images is data from a single power-on cycle. Therefore, scene-based statistical radiometric calibration methods cannot meet the radiometric calibration accuracy requirements for very high-sensitivity infrared imaging data of the ocean. Summary of the Invention

[0006] The technical problem solved by this invention is to overcome the shortcomings of the prior art and provide a high-precision radiometric calibration method for very sensitive infrared imaging data of the sea. It combines the radiometric calibration source method and the scene calibration method. Based on the radiometric calibration source correction, it uses a small sample scene data statistical method to further improve the radiometric calibration accuracy.

[0007] The technical solution of this invention is: a high-precision radiometric calibration method for very sensitive infrared imaging data of the sea, comprising:

[0008] Step 1: Read in the on-board blackbody calibration data, and perform blind element detection and calculate the relative radiation correction coefficient based on the blackbody calibration data;

[0009] Step 2: Read in the infrared sea imaging data and the relative radiometric correction coefficient, and perform initial relative radiometric correction on the sea imaging image;

[0010] Step 3: Read in all images after the initial relative radiometric correction, perform statistical correction, and remove residual stripe noise;

[0011] Step 4: Perform flash element detection and removal on the statistically corrected image to obtain the final output image.

[0012] Furthermore, in step 1, the blind element detection criteria include: the blackbody calibration data response is 0 or the response is saturated; the noise is greater than twice the mean noise.

[0013] Furthermore, in step 2, the coefficients calculated based on the blackbody calibration data are initially corrected, including relative radiation correction and blind element replacement;

[0014] The relative correction includes relative radiometric correction based on calibration factors:

[0015] F(i,j)=G(i,j)×K(i)+Offset(i)

[0016] In the formula, G(i,j) is the original response of the i-th pixel in the j-th sampling; F(i,j) is the corrected response of the i-th pixel in the j-th sampling; K(i) is the gain value of the relative radiometric correction coefficient of the i-th pixel; and Offset(i) is the offset value of the relative radiometric correction coefficient of the i-th pixel.

[0017] The blind cell compensation includes: if the i-th cell is a blind cell, then F′(i,j)=F(i-1,j) / 2+F(i+1,j) / 2 is used to replace the blind cell; where F′(i,j) is the response of the i-th cell after the j-th blind cell removal.

[0018] Furthermore, in step 3, statistical correction is performed to remove residual stripe noise, as follows:

[0019] Based on the sea imaging data after initial relative radiometric correction, the histogram curve of each pixel is calculated.

[0020] Based on the sea imaging data obtained in the previous step after initial relative radiometric correction, the standard histogram curves of all pixels are calculated.

[0021] Based on the histogram curve of each pixel, calculate the normalized cumulative histogram curve of the pixels of that pixel.

[0022] Based on the standard histogram curves of all pixels, calculate the standard normalized cumulative histogram curves of all pixels.

[0023] Based on the normalized cumulative histogram curve for each pixel, the DN values ​​corresponding to the curve at 0.05 and 0.95 are calculated as DN. mini and DN maxi ;

[0024] Based on the standard normalized cumulative histogram curves of all detector elements, the DN values ​​corresponding to the curves at 0.05 and 0.95 are calculated as DN... min and DN max ;

[0025] Based on the normalized cumulative histogram curve obtained in the previous step, a radiation correction lookup table is constructed. Make K′=DN min ,DN min +1,DN min +2,…,DN max ;

[0026] Using the lookup table obtained in the previous step, the image F′ after the initial radiometric correction is radiometrically corrected again to remove residual stripe noise in the image and obtain the corrected image F″.

[0027] Furthermore, in step 4, flash element detection and removal includes:

[0028] The difference between odd and even elements is calculated based on the image F″, including the even element difference response C1 and the odd element difference response C2.

[0029] The positions where the difference between even and odd elements is greater than a threshold S are extracted based on the even element difference response C1 and the odd element difference response C2.

[0030] Based on the position where the difference between odd and even elements is greater than a threshold, determine whether it is a flash element according to preset judgment conditions;

[0031] Based on the detected flash elements, flash elements are removed using the same method as blind elements.

[0032] Furthermore, the preset determination condition is: if a certain element has more than 50% parity-even difference in all samples of an image, then this element is identified as a flash element.

[0033] A high-precision radiometric calibration system for very sensitive infrared ocean imaging data includes:

[0034] The first module reads in the on-board blackbody calibration data and performs blind element detection and calculates the relative radiation correction coefficient based on the blackbody calibration data.

[0035] The second module reads in the infrared sea imaging data and the relative radiometric correction coefficient, and performs initial relative radiometric correction on the sea imaging image.

[0036] The third module reads in all images after the initial relative radiometric correction, performs statistical correction, and removes residual stripe noise.

[0037] The fourth module performs flash pixel detection and removal on the statistically corrected image to obtain the final output image.

[0038] Furthermore, in the first module, the blind element detection criteria include: the blackbody calibration data response is 0 or the response is saturated; the noise is greater than twice the mean noise.

[0039] In the second module, the coefficients calculated based on the blackbody calibration data are initially corrected, including relative radiometric correction and blind element replacement;

[0040] The relative correction includes relative radiometric correction based on calibration factors:

[0041] F(i,j)=G(i,j)×K(i)+Offset(i)

[0042] In the formula, G(i,j) is the original response of the i-th pixel in the j-th sampling; F(i,j) is the corrected response of the i-th pixel in the j-th sampling; K(i) is the gain value of the relative radiometric correction coefficient of the i-th pixel; and Offset(i) is the offset value of the relative radiometric correction coefficient of the i-th pixel.

[0043] The blind cell compensation includes: if the i-th cell is a blind cell, then F′(i,j)=F(i-1,j) / 2+F(i+1,j) / 2 is used to replace the blind cell; where F′(i,j) is the response of the i-th cell after the j-th blind cell removal;

[0044] In the third module, statistical correction is performed to remove residual stripe noise, as detailed below:

[0045] Based on the sea imaging data after initial relative radiometric correction, the histogram curve of each pixel is calculated.

[0046] Based on the sea imaging data obtained in the previous step after initial relative radiometric correction, the standard histogram curves of all pixels are calculated.

[0047] Based on the histogram curve of each pixel, calculate the normalized cumulative histogram curve of the pixels of that pixel.

[0048] Based on the standard histogram curves of all pixels, calculate the standard normalized cumulative histogram curves of all pixels.

[0049] Based on the normalized cumulative histogram curve for each pixel, the DN values ​​corresponding to the curve at 0.05 and 0.95 are calculated as DN. mini and DN maxi ;

[0050] Based on the standard normalized cumulative histogram curves of all detector elements, the DN values ​​corresponding to the curves at 0.05 and 0.95 are calculated as DN... min and DN max ;

[0051] Based on the normalized cumulative histogram curve obtained in the previous step, a radiation correction lookup table is constructed. Make K′=DN min ,DN min +1,DN min +2,…,DN max ;

[0052] The lookup table obtained in the previous step is used to perform a second radiometric correction on the image F′ after the initial radiometric correction, and the residual stripe noise in the image is removed to obtain the corrected image F″.

[0053] The fourth module includes flash element detection and removal:

[0054] The difference between odd and even elements is calculated based on the image F″, including the even element difference response C1 and the odd element difference response C2.

[0055] The positions where the difference between even and odd elements is greater than a threshold S are extracted based on the even element difference response C1 and the odd element difference response C2.

[0056] Based on the position where the difference between odd and even elements is greater than a threshold, determine whether it is a flash element according to preset judgment conditions;

[0057] Based on the detected flash elements, flash elements are removed using the same method as blind elements.

[0058] The preset determination condition is: if a certain element has more than 50% difference between odd and even elements in all samples of an image, then this element is identified as a flash element.

[0059] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of a high-precision radiometric calibration method for very sensitive infrared imaging data of the sea.

[0060] A high-precision radiometric calibration device for very-sensitive infrared sea imaging data includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the high-precision radiometric calibration method for very-sensitive infrared sea imaging data.

[0061] The advantages of this invention compared to the prior art are:

[0062] (1) Blind element detection is no longer based on the national standard, but on the improved standard. Compared with the national standard, the number of blind elements is reduced, but the calibration accuracy remains unchanged.

[0063] (2) The radiation calibration accuracy is high, and the radiation calibration accuracy has been improved from the original 3% to 1%;

[0064] (3) No longer relying on large sample data, the radiation calibration coefficients are obtained based on small sample data;

[0065] (4) Real-time monitoring and processing of flash elements were considered to further improve the radiation quality of infrared images. Attached Figure Description

[0066] Figure 1 This is a flowchart of the method of the present invention.

[0067] Figure 2 This is a schematic diagram of the flash element of the present invention. Detailed Implementation

[0068] To better understand the above technical solutions, the technical solutions of this application will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of this application and the specific features in the embodiments are detailed descriptions of the technical solutions of this application, rather than limitations on the technical solutions of this application. In the absence of conflict, the embodiments of this application and the technical features in the embodiments can be combined with each other.

[0069] The following description, in conjunction with the accompanying drawings, provides a more detailed explanation of a high-precision radiometric calibration method for very sensitive infrared ocean imaging data provided in this application. Specific implementation methods may include (e.g.) Figures 1-2 As shown):

[0070] Step 1: Read in the on-board blackbody calibration data, and perform blind element detection and calculate the relative radiation correction coefficient based on the blackbody calibration data;

[0071] Step 2: Read in the infrared sea imaging data and the relative radiometric correction coefficient, and perform initial relative radiometric correction on the sea imaging image;

[0072] Step 3: Read in all images after the initial relative radiometric correction, perform statistical correction, and remove residual stripe noise;

[0073] Step 4: Perform flash element detection and removal on the statistically corrected image to obtain the final output image.

[0074] Furthermore, in step 1, the blind element detection criteria include: the blackbody calibration data response is 0 or the response is saturated; the noise is greater than twice the mean noise.

[0075] In one possible implementation, in step 2, the coefficients calculated based on the blackbody calibration data are initially corrected, including relative radiometric correction and blind element replacement.

[0076] The relative correction includes relative radiometric correction based on calibration factors:

[0077] F(i,j)=G(i,j)×K(i)+Offset(i)

[0078] In the formula, G(i,j) is the original response of the i-th pixel in the j-th sampling; F(i,j) is the corrected response of the i-th pixel in the j-th sampling; K(i) is the gain value of the relative radiometric correction coefficient of the i-th pixel; and Offset(i) is the offset value of the relative radiometric correction coefficient of the i-th pixel.

[0079] Optionally, in one possible implementation, the blind cell compensation includes: if the i-th cell is a blind cell, then F′(i,j) = F(i-1,j) / 2 + F(i+1,j) / 2 is used to replace the blind cell; where F′(i,j) is the response of the i-th cell after the j-th blind cell removal.

[0080] In one possible implementation, step 3 involves statistical correction to remove residual stripe noise, specifically as follows:

[0081] Based on the sea imaging data after initial relative radiometric correction, the histogram curve of each pixel is calculated.

[0082] Based on the sea imaging data obtained in the previous step after initial relative radiometric correction, the standard histogram curves of all pixels are calculated.

[0083] Based on the histogram curve of each pixel, calculate the normalized cumulative histogram curve of the pixels of that pixel.

[0084] Based on the standard histogram curves of all pixels, calculate the standard normalized cumulative histogram curves of all pixels.

[0085] Based on the normalized cumulative histogram curve for each pixel, the DN values ​​corresponding to the curve at 0.05 and 0.95 are calculated as DN. mini and DN maxi ;

[0086] Based on the standard normalized cumulative histogram curves of all detector elements, the DN values ​​corresponding to the curves at 0.05 and 0.95 are calculated as DN... min and DN max ;

[0087] Based on the normalized cumulative histogram curve obtained in the previous step, a radiation correction lookup table is constructed. Make K′=DN min ,DN min +1,DN min +2,…,DN max ;

[0088] Using the lookup table obtained in the previous step, the image F′ after the initial radiometric correction is radiometrically corrected again to remove residual stripe noise in the image and obtain the corrected image F″.

[0089] Furthermore, in one possible implementation, step 4, flash element detection and removal, includes:

[0090] The difference between odd and even elements is calculated based on the image F″, including the even element difference response C1 and the odd element difference response C2.

[0091] The positions where the difference between even and odd elements is greater than a threshold S are extracted based on the even element difference response C1 and the odd element difference response C2.

[0092] Based on the position where the difference between odd and even elements is greater than a threshold, determine whether it is a flash element according to preset judgment conditions;

[0093] Based on the detected flash elements, flash elements are removed using the same method as blind elements.

[0094] Furthermore, the preset determination condition is: if a certain element has more than 50% parity-even difference in all samples of an image, then this element is identified as a flash element.

[0095] Based on and Figure 1 With the same inventive concept, the present invention also provides a high-precision radiometric calibration system for very sensitive infrared imaging data of the sea, comprising:

[0096] The first module reads in the on-board blackbody calibration data and performs blind element detection and calculates the relative radiation correction coefficient based on the blackbody calibration data.

[0097] The second module reads in the infrared sea imaging data and the relative radiometric correction coefficient, and performs initial relative radiometric correction on the sea imaging image.

[0098] The third module reads in all images after the initial relative radiometric correction, performs statistical correction, and removes residual stripe noise.

[0099] The fourth module performs flash pixel detection and removal on the statistically corrected image to obtain the final output image.

[0100] The solution provided in the embodiments of this application specifically includes:

[0101] (1) Blind element detection and relative radiation correction coefficient calculation based on on-board blackbody calibration data

[0102] The national standard for blind pixel testing defines pixels as follows: pixels with a response slope less than half the mean response slope and pixels with noise greater than twice the mean noise. Experiments have shown that some pixels with a response slope less than half the mean response slope perform normally after radiometric correction. Therefore, the blind pixel testing standard is as follows:

[0103] ① The calibrated blackbody response is 0 or the response is saturated;

[0104] ② The noise level is more than twice the mean noise level.

[0105] The formula for calculating the coefficient is as follows:

[0106]

[0107]

[0108] In the formula,

[0109] K(i) and Offset(i) are the gain and offset of the correction coefficient for the i-th pixel, i = 1, 2, 3, ..., N;

[0110] The mean response of high and low temperature blackbody calibration data (excluding blind elements);

[0111] DN h (i), DN l (i) represents the response of the high and low temperature blackbody calibration data of the i-th pixel.

[0112] (2) Correct the image based on the relative radiometric correction coefficients calculated in the previous step.

[0113] (2.1) Using the calibration coefficients obtained in the previous step, perform relative radiation correction using the following formula:

[0114] F(i,j)=G(i,j)×K(i)+Offset(i) (3)

[0115] In the formula,

[0116] G(i,j) is the original response of the i-th pixel in the j-th sampling;

[0117] F(i,j) is the response of the i-th pixel after the j-th sampling correction.

[0118] (2.2) Blind Pixel Compensation

[0119] If the i-th element is a blind element, then the blind element replacement is performed using the following formula:

[0120] F′(i,j)=F(i-1,j) / 2+F(i+1,j) / 2 (4)

[0121] (3) Statistical correction based on small sample data

[0122] Histogram statistics are not suitable for small sample data because the histogram distribution of each pixel in a small sample differs from the histogram distribution of all pixels. For sea surface infrared remote sensing data, the histogram is mainly represented at both ends. At the lower end of the histogram, clouds (clouds tend to have lower temperatures) are the main features, while at the higher end, ships or small islands are the main targets. Therefore, the original histogram statistics are improved as follows:

[0123] (3.1) For the image F′ obtained in the previous step, calculate the histogram curve of each data element;

[0124]

[0125] In the formula:

[0126] Q(i,k) is the probability of the sample with DN response k in the i-th pixel;

[0127] N k Let k be the total number of times the sample with DN response k occurs in the i-th pixel;

[0128] N A The total number of samples for each cell;

[0129] (3.2) For the image F′ obtained in the previous step, calculate the standard histogram curve of all pixels based on the following formula.

[0130]

[0131] (3.3) Calculate the normalized cumulative histogram H(i,K′) of the i-th probe using the following formula:

[0132]

[0133] 0≤H(i,K′)≤1

[0134] (3.4) Calculate the standard normalized cumulative histogram of all detector elements using the following formula:

[0135]

[0136] 0≤Z′(K′)≤1

[0137] (3.5) Based on the normalized cumulative histogram H(i,K′), calculate the DN corresponding to H(i,K′)=0.05 and H(i,K′)=0.95. mini and DN maxi ;

[0138] (3.6) Based on the standard normalized cumulative histogram Z′(K′) of all detector elements, calculate the DN corresponding to Z′(K′) = 0.05 and Z′(K′) = 0.95. min and DN max ;

[0139] (3.7) Constructing a radiation correction lookup table Make K′=DN min ,DN min +1,DN min +2,…,DN max .

[0140] (3.8) Use the lookup table obtained in the previous step to perform radiometric correction on F′ to obtain the corrected image F″.

[0141] (4) Flash element detection and removal

[0142] During processing, occasional blind pixel noise was found in the infrared images, as shown in the figure below. Pixels with this type of noise sometimes respond normally and sometimes abnormally. If blind pixel noise is removed according to fixed locations, ground feature information will be lost. Therefore, it is necessary to detect and remove this type of noise in real time.

[0143] (4.1) Calculate the difference between odd and even elements

[0144] The mean response G1 for even rows and the mean response G2 for odd rows are calculated using formulas (1) and (2), and the difference response C1 for even rows and the difference response C2 for odd rows are calculated based on formulas (3) and (4).

[0145] G1(i,2×j)=F″(i,2×j-1) / 2+F″(i,2×j+1) / 2 (1)

[0146] G1(i,2×j+1)=F″(i,2×j) / 2+F″(i,2×j+2) / 2 (2)

[0147] C1=|G1-F″| (3)

[0148] C2=|G2-F″| (4)

[0149] (4.2) Extracting positions with large differences between odd and even elements

[0150] Based on the even-numbered element difference response C1 and the odd-numbered element difference response C2 obtained in the previous step, the positions where the odd-even element difference is greater than the threshold S are extracted according to formula (5).

[0151]

[0152] in,

[0153] S = max(S1, S2),

[0154] (4.3) Filter out the location of the flash element

[0155] Analysis of flash pixels revealed significant differences between them and their neighboring pixels, allowing for extraction using a difference detection method. The specific extraction method is shown in formula (6). If a pixel exhibits more than 50% odd-even pixel differences across all samples of the image, it can be identified as a flash pixel.

[0156]

[0157] in,

[0158] (4.4) Flash element removal

[0159] The method for removing flash elements is the same as the method for removing blind elements.

[0160] This application provides a computer-readable storage medium storing computer instructions that, when executed on a computer, cause the computer to perform... Figure 1 The method described.

[0161] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0162] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0163] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0164] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0165] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

[0166] The contents not described in detail in this specification are common knowledge to those skilled in the art.

Claims

1. A method for high-precision radiometric calibration of very high sensitivity infrared against-sea imaging data, characterized in that, include: Step 1: Read in the on-board blackbody calibration data, and perform blind element detection and calculate the relative radiation correction coefficient based on the blackbody calibration data; Step 2: Read in the infrared sea imaging data and the relative radiometric correction coefficient, and perform initial relative radiometric correction on the sea imaging image; Step 3: Read in all images after the initial relative radiometric correction, perform statistical correction, and remove residual stripe noise; Step 4: Perform flash element detection and removal on the statistically corrected image to obtain the final output image; In step 3, statistical correction is performed to remove residual stripe noise, as follows: Based on the sea imaging data after initial relative radiometric correction, the histogram curve of each pixel is calculated. Based on the sea imaging data obtained in the previous step after initial relative radiometric correction, the standard histogram curves of all pixels are calculated. Based on the histogram curve of each pixel, calculate the normalized cumulative histogram curve of that pixel. Based on the standard histogram curves of all pixels, calculate the standard normalized cumulative histogram curves of all pixels. Based on the normalized cumulative histogram curve of each pixel, the DN values corresponding to 0.05 and 0.95 of the curve are calculated respectively as and ; Based on the standard normalized cumulative histogram curve of all the probe elements, the DN values corresponding to 0.05 and 0.95 of the curve are calculated as DN1 and DN2, respectively and ; Based on the normalized cumulative histogram curve obtained in the previous step, a radiation correction lookup table is constructed. , making = , ; using the look-up table obtained in the previous step performing a second radiation correction to remove residual stripe noise from the image to obtain a corrected image .

2. The method of claim 1, wherein the method is characterized by: In step 1, the blind element detection criteria include: the blackbody calibration data response is 0 or the response is saturated; the noise is greater than twice the mean noise.

3. The method of claim 1, wherein the method is characterized by: In step 2, the coefficients calculated based on the blackbody calibration data are initially corrected, including relative radiation correction and blind element replacement. Relative correction includes relative radiometric correction based on calibration factors: In the formula, is the original response of the i-th pixel in the j-th sampling; is the corrected response of the i-th pixel in the j-th sampling, is the relative radiation correction coefficient gain value of the i-th pixel, is the relative radiation correction coefficient offset value of the i-th pixel; Blind cell compensation includes: if the i-th cell is a blind cell, then carrying out blind cell replacement; wherein, is the response of the i-th pixel after the j-th blind cell removal.

4. The method of claim 1, wherein the method is a high-precision radiometric calibration method for very-high-sensitivity infrared sea-surface imaging data. Step 4, flash element detection and removal, includes: Based on images Computing parity element differences, including even element difference responses , odd element difference responses ; Even element difference value response and odd element difference value response extracting positions where the odd-even element difference is greater than a threshold S Based on the position where the difference between odd and even elements is greater than a threshold, determine whether it is a flash element according to preset judgment conditions; Based on the detected flash elements, flash elements are removed using the same method as blind elements.

5. The method of claim 4, wherein the method further comprises: The preset determination condition is: if a certain element has more than 50% difference between odd and even elements in all samples of an image, then this element is identified as a flash element.

6. A high precision radiometric calibration system for very high sensitivity infrared against sea imaging data, characterized in that, include: The first module reads in the on-board blackbody calibration data and performs blind element detection and calculates the relative radiation correction coefficient based on the blackbody calibration data. The second module reads in the infrared sea imaging data and the relative radiometric correction coefficient, and performs initial relative radiometric correction on the sea imaging image. The third module reads in all images after the initial relative radiometric correction, performs statistical correction, and removes residual stripe noise. The fourth module performs flash pixel detection and removal on the statistically corrected image to obtain the final output image; In the third module, statistical correction is performed to remove residual stripe noise, as detailed below: Based on the sea imaging data after initial relative radiometric correction, the histogram curve of each pixel is calculated. Based on the sea imaging data obtained in the previous step after initial relative radiometric correction, the standard histogram curves of all pixels are calculated. Based on the histogram curve of each pixel, calculate the normalized cumulative histogram curve of that pixel. Based on the standard histogram curves of all pixels, calculate the standard normalized cumulative histogram curves of all pixels. Based on the normalized cumulative histogram curve of each pixel, the DN values corresponding to 0.05 and 0.95 of the curve are calculated respectively as and ; Based on the standard normalized cumulative histogram curve of all the probe elements, the DN values corresponding to 0.05 and 0.95 of the curve are calculated as DN1 and DN2, respectively and ; Based on the normalized cumulative histogram curve obtained in the previous step, a radiation correction lookup table is constructed , such that = , ; using the look-up table obtained in the previous step performing a second radiation correction to remove residual stripe noise from the image to obtain a corrected image .

7. A high precision radiometric calibration system for very high sensitivity infrared against sea imaging data according to claim 6, characterized in that, In the first module, the blind element detection criteria include: the blackbody calibration data response is 0 or the response is saturated; the noise is greater than twice the mean noise. In the second module, the coefficients calculated based on the blackbody calibration data are initially corrected, including relative radiometric correction and blind element replacement; Relative correction includes relative radiometric correction based on calibration factors: In the formula, This represents the original response of the i-th pixel during the j-th sampling. The response of the i-th pixel after the j-th sampling correction is... Let be the relative radiometric correction coefficient gain value for the i-th pixel. This is the offset value of the relative radiometric correction coefficient for the i-th pixel; Blind cell compensation includes: if the ith cell is a blind cell, then performing blind cell replacement; wherein, is the response of the ith pixel after the jth blind cell removal. The fourth module includes flash element detection and removal: Based on image Computing parity element difference, including even element difference response , odd element difference response ; Even element difference value response and odd element difference value response extracting positions where the odd-even element difference is greater than a threshold S Based on the position where the difference between odd and even elements is greater than a threshold, determine whether it is a flash element according to preset judgment conditions; Based on the detected flash elements, flash elements are removed using the same method as blind elements. The preset determination condition is: if a certain element has more than 50% difference between odd and even elements in all samples of an image, then this element is identified as a flash element.

8. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1-7. When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 5.

9. A high-precision radiometric calibration device for very high sensitivity infrared against sea imaging data, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 5.

Images