Dynamic scene infrared image pot cover removing method

Abstract

The invention belongs to the technical field of infrared image processing, and particularly relates to a dynamic scene infrared image pot cover removing method. The method comprises the following steps: an infrared thermal imaging device closes a pot cover removing algorithm by default, inputs an infrared original image, and calculates and models after receiving a pot cover obtaining instruction to obtain a pot cover parameter table DistTable; calculating the distance from each pixel point to the center of the image, and querying a pot cover parameter table DistTable according to the distances to obtain a pot cover correction value; and subtracting the pot cover correction value from each pixel point to obtain image data after pot cover removal correction. The method for removing the pot cover from the infrared image in the dynamic scene is realized in real time, and the infrared imaging system can provide more stable and higher-quality image output under various environmental conditions.

Description

Technical Field

[0001] The invention belongs to the technical field of infrared image processing, and in particular relates to a method for removing pot covers from dynamic scene infrared images. Background Art

[0002] The pot lid effect of infrared images is caused by uneven energy reception between the center and surrounding areas of the detector due to the thermal radiation of the lens structure and uneven optical energy of the lens, resulting in uneven imaging brightness. Characteristics of the pot lid effect: First, the pot lid phenomenon will continue to deepen as the machine cold starts to thermal equilibrium; second, when the thermal equilibrium changes with the ambient temperature, the pot lid phenomenon will change accordingly. The pot lid effect will cause the image to be white all around and black in the middle, and the image layering and details will be lost, affecting the image visual effect and subsequent processing and analysis. To solve the pot lid effect of the image, the structural heat dissipation design must be considered first. If the heat dissipation cannot be solved, a pot lid removal algorithm needs to be added.

[0003] There are two types of pot lid removal algorithms: calibration method and real-time calculation method. The calibration method is to collect uniform surface images and calculate compensation templates, and then use the templates to perform real-time pot lid removal compensation on the images; after calibration, the template image can be directly subtracted as a correction.

[0004] Mathematical modeling can also be performed, and real-time calculation and compensation can be performed. This algorithm basically solves the impact of the pot cover on the image. In order to adapt to environmental changes, multiple pot cover templates are collected and stored at different ambient temperatures. When in use, the compensation coefficient of the pot cover can be calculated in real time through the temperature sensing module inside the movement, and the corresponding pot cover template can be selected to realize the removal of the pot cover from the cold machine to the hot machine. Consumes computing time and storage resources. The algorithm for real-time calculation and compensation does not require data calibration. Generally, distance function or block fitting is used for mathematical modeling, and then the image is processed to remove the pot cover.

[0005] Patent CN117911290A introduces an infrared image pot lid correction algorithm, which processes the infrared image of the pot lid in a grid-like manner and establishes a grid correction coefficient table; processes the infrared image containing the pot lid phenomenon in blocks, calculates the grid associated with the block to which the pixel to be corrected belongs, and obtains the grid correction coefficient of the corresponding grid from the grid correction coefficient table; calculates the actual correction coefficient of the pixel to be corrected by the grid correction coefficient related to the pixel to be corrected; and multiplies the pixel value of the pixel to be corrected by the actual correction coefficient to obtain the correction value of the pixel to be corrected. The shortcomings of the method are: (1) Calibration is required to collect template data; (2) When the template collection environment is different from the use environment, it will cause over-correction or under-correction. (3) From the analysis of the cause of pot lid formation, the pot lid effect is a low-frequency additive noise. After mathematical modeling, this patent uses multiplication operation for correction, which will cause over-correction and affect the scene effect.

[0006] Patent CN113421220A provides a method for removing pot lids from infrared images, including: a calibration process and an actual practical process; in the calibration process, a cubic polynomial function between the singular value of an ideal pot-free infrared image and the infrared focal plane temperature ti and a quadratic function to the image pic1 and the infrared focal plane temperature ti are obtained; in the actual practical process, the cubic function and quadratic function of the calibration process are used to remove the pot lid of the infrared image in the actual scene. The shortcomings of the method are: (1) The calibration process is cumbersome. (2) The use process still requires the acquisition of a uniform surface infrared image, and then the estimated background is obtained according to the fitting calculation, multiplied by the K coefficient, to obtain a pot-free image. This uniform surface infrared image is obtained through a shutter plate, which is generally installed between the lens and the detector, and cannot remove the pot lid caused by the temperature rise of the lens barrel.

[0007] Patent CN117197034A discloses a method, storage medium and device for suppressing black pots in infrared images. The method comprises: calculating the distance from the pixel point of the original image to the center of the black pot; the distance between the pixel point of the original image and the center of the black pot is obtained by calculating the difference between the horizontal and vertical coordinates according to the coordinates of the pixel point and the coordinates of the center of the black pot; calculating the edge coefficient of each pixel point of the original image according to the distance between the pixel point and the center of the black pot; the edge coefficient is obtained by mapping the distance between the pixel point and the center of the black pot and multiplying it by an adjustment parameter; and calculating a pot lid suppression image according to the edge coefficient and the original image. The method makes full use of factors such as the causes and effects of the pot lid effect, maps the distance to the edge coefficient for compensation, does not require the pre-storage of a large amount of data, and is more suitable for engineering applications. The shortcomings of the method are as follows: (1) No pre-calibration is required, and the pot lid template correction value is obtained by calculating the distance function, which requires manual parameter adjustment and is not automatically corrected with the environment. (2) The edge coefficient matrix corrects each pixel value by point-to-point multiplication, and removing the pot lid will also lose scene details. Summary of the invention

[0008] To this end, the present invention proposes a method for removing pot lids from dynamic scene infrared images, which does not require collecting pot lid templates at multiple temperature points, and performs modeling and correction according to the environmental conditions during use to adapt to various scenarios.

[0009] The present invention proposes a method for removing pot lids from dynamic scene infrared images, which can be used to calibrate, model and correct the pot lids of infrared thermal imaging equipment in use. The method can be used immediately after calculation, and the current real-time video can be corrected with the currently calculated pot lid data, and the pot lid noise can be removed as low-frequency false noise. The key is to remove the high-frequency information in the scene without affecting the calculation of pot lid parameters, and to avoid ghosting caused by scene imaging.

[0010] The steps of the pot cover removal method of the present invention are described as follows: (1) The infrared thermal imaging device turns off the pot cover removal algorithm by default. After inputting the original infrared image and receiving the pot cover acquisition instruction, the model is calculated and modeled to obtain the pot cover parameter table DistTable; (2) Calculate the distance from each pixel to the center of the image, query the pot cover parameter table DistTable based on the distance, and obtain the pot cover correction value; (3) Subtract the pot lid correction value from each pixel to obtain the image data after removing the pot lid correction.

[0011] After receiving the "get pot lid" command, the infrared thermal imaging device calculates and models the pot lid parameter table, which will be cleared when the power is off. Send the "get pot lid" command again, calculate the pot lid parameter table with the current scene image, and replace the previous pot lid parameter table.

[0012] The steps for computational modeling and obtaining the pot lid parameter table are as follows: (1) Input the current frame scene image SrcImage, perform Gaussian filtering, and obtain GaussFilterImage; (2) Perform logarithmic transformation on the Gaussian filtered image and convert it to the log domain to obtain GaussLogImage; (3) Gradient operation, extracting the horizontal gradient matrix GradX and vertical gradient matrix GradY of GaussLogImage; (4) Convert the matrices GradX and GradY into one-dimensional arrays respectively and concatenate them to obtain a one-dimensional array Grad with twice the image resolution; (5) Calculate the coordinate position correlation matrix for each pixel point, Q in the horizontal direction and R in the vertical direction. The parameter N controls the dimension nDim of the correlation matrix. ; (6) Convert the Q and R array matrices of each pixel into one-dimensional arrays Mx and My of five elements, and concatenate and merge them to obtain the position matrix M; (7) Calculate the Gamma parameter table ; Gamma expansion and 0 padding are convenient for the next step: ; (8) Calculate the bias matrix Bias: Traverse each pixel of the image, calculate the bias parameter value Sum, and obtain the bias matrix Bias: ; (9) Bias correction: convert the original image to the logarithmic domain and then perform subtraction ; (10) Exponential transformation to obtain the corrected image: ; (11) Calculate the pot cover matrix template GGMask: ; (12) Divide the pot cover template into a distance ring RingImg with the center pixel of the image as the center according to the distance difference; (13) Calculate the pixel mean of each ring and generate the mean list array RingGrayMean of the distance to the center of the image; (14) Perform neighborhood smoothing correction on the distance list array RingGrayMean to generate a new mean list RingGrayMeanSmooth; (15) Distance-weighted interpolation is performed on the mean list RingGrayMeanSmooth, and the distance of each pixel in the image is included to obtain the final pot cover distance parameter table DistTable.

[0013] As in the above steps (1~15), the pot cover modeling process is completed. Then, the DistTable can be used to calculate and calibrate the pot cover. The pot cover modeling process does not require real-time calculation. Receive the "Get Pot Cover" command and execute it once. Use DistTable for real-time correction. It will be cleared when the power is off. Send the "Get Pot Cover" command again, calculate the pot cover parameter table once with the current scene image, and replace the last pot cover parameter table.

[0014] The method for removing pot lids from dynamic scene infrared images of the present invention is implemented in real time, and the infrared imaging system can provide more stable and higher quality image output under various environmental conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 The figure is a flow chart of the method of the present invention.

[0016] Figure 2 Modeling flow chart of pot cover parameter table.

[0017] Figure 3 Input the original pot cover scene image.

[0018] Figure 4 Calculate and obtain the pot cover template GGMask.

[0019] Figure 5 Schematic diagram of the distance ring.

[0020] Figure 6 Distance ring mean table curve.

[0021] Figure 7 Pot lid correction image. DETAILED DESCRIPTION

[0022] The present invention introduces a method for removing pot lids from dynamic scene infrared images, and performs pot lid removal calibration, modeling and correction on an uncooled infrared thermal imaging module device with a resolution of 1280*1024 that is in use.

[0023] The steps for removing the lid are as follows: (1) Input the original infrared image and enable the pot removal algorithm. The pot removal algorithm is disabled by default.

[0024] (2) Calculate the distance from each pixel to the center of the image, query the pot cover parameter table DistTable based on the distance, and obtain the pot cover correction value. The table length is 1 / 2 of the image diagonal and does not exceed 820.

[0025] (3) Subtract the pot lid correction value from each pixel to obtain the image data after removing the pot lid correction.

[0026] After receiving the "get pot lid" command, the infrared thermal imaging device calculates and models the pot lid parameter table, which will be cleared when the power is off. Send the "get pot lid" command again, calculate the pot lid parameter table with the current scene image, and replace the previous pot lid parameter table.

[0027] The steps to model the pot cover and obtain the pot cover parameter table are as follows: (1) Input the current frame scene image SrcImage, perform Gaussian filtering, and obtain GaussFilterImage.

[0028] (2) Perform a logarithmic transformation on the Gaussian filtered image and convert it to the log domain to obtain GaussLogImage.

[0029] (3) Gradient operation, extracting the horizontal gradient matrix GradX and vertical gradient matrix GradY of GaussLogImage.

[0030] (4) Convert the matrices GradX and GradY into one-dimensional arrays respectively and concatenate them to obtain an array Grad with twice the image resolution.

[0031] (5) Calculate the coordinate position correlation matrix for each pixel point, Q in the horizontal direction and R in the vertical direction. Parameter N is the dimension parameter, with the default value of 2, which controls the dimension nDim of the correlation matrix; ;

[0032] i, j are dimension loop variables, x represents the vertical coordinate, and y represents the horizontal coordinate.

[0033] (6) Convert the Q and R array matrices of each pixel into one-dimensional arrays Mx and My of five elements, and concatenate them to obtain the position matrix M.

[0034] (7) Calculate the Gamma parameter table.

[0035] ;

[0036] Gamma expansion is padded with 0 for the convenience of the next step.

[0037] ; (8) Calculate the bias matrix Bias. Traverse each pixel of the image, calculate the bias parameter value, and obtain the bias matrix Bias.

[0038] ; (9) Bias correction: The original image is converted to the logarithmic domain and then subtracted.

[0039] ; (10) Exponential transformation to obtain the corrected image.

[0040] ; (11) Calculate the pot cover matrix template GGMask.

[0041] ; (12) The pot cover template is divided into a distance ring RingImg with the center pixel of the image as the center according to the distance difference, and the width of the ring is 20. (13) Calculate the pixel mean of each ring and generate an array RingGrayMean of mean lists of distances to the center of the image.

[0042] (14) Perform neighborhood smoothing correction on the distance list array RingGrayMean to generate a new mean list RingGrayMeanSmooth.

[0043] (15) Distance-weighted interpolation is performed on the mean list RingGrayMeanSmooth, and the distance of each pixel in the image is included to obtain the final pot cover distance parameter table DistTable.

[0044] As in the above steps (1-15), the pot lid modeling process is completed. Then, the pot lid calculation and correction can be performed using DistTable. The pot lid modeling process does not require real-time calculation. It receives the "Get Pot Lid" command and executes it once. Real-time Correction Using DistTable By using the real-time implementation technology of the pot cover removal method of the present invention, the infrared imaging system can provide more stable and higher quality image output under various environmental conditions.

[0045] The method for removing pot lids from infrared images disclosed in the present invention is based on real-time calibration, modeling, and correction of scenes. That is, when a pot lid effect appears during the operation of an infrared camera, a pot lid model is modeled based on the current environment and scene, and pot lid removal correction is performed on real-time video frames. There is no need for multi-temperature point calibration, a large number of pot lid model storage, and it can be implemented on an embedded chip, thereby improving the quality of infrared images. There is no need to adjust the parameters of the pot lid removal method. Before the infrared camera leaves the factory, there is no need to calibrate and store the pot lid removal matrix template; during the use of the infrared camera, there is no need to calibrate and store the pot lid removal matrix template for the temperature uniform surface.

Claims

1. A method for removing pot lids from dynamic scene infrared images, characterized in that: The steps include: (1) The infrared thermal imaging device turns off the pot cover removal algorithm by default. After inputting the original infrared image and receiving the pot cover acquisition instruction, the model is calculated and modeled to obtain the pot cover parameter table DistTable; (2) Calculate the distance from each pixel to the center of the image, query the pot cover parameter table DistTable based on the distance, and obtain the pot cover correction value; (3) Subtract the pot lid correction value from each pixel to obtain the image data after removing the pot lid correction.

2. The method for removing pot lid from dynamic scene infrared image according to claim 1, characterized in that: The steps for computational modeling and obtaining the pot lid parameter table are as follows: (1) Input the current frame scene image SrcImage, perform Gaussian filtering, and obtain GaussFilterImage; (2) Perform logarithmic transformation on the Gaussian filtered image and convert it to the log domain to obtain GaussLogImage; (3) Gradient operation, extracting the horizontal gradient matrix GradX and vertical gradient matrix GradY of GaussLogImage; (4) Convert the matrices GradX and GradY into one-dimensional arrays respectively and concatenate them to obtain a one-dimensional array Grad with twice the image resolution; (5) Calculate the coordinate position correlation matrix for each pixel point, Q in the horizontal direction and R in the vertical direction. The parameter N controls the dimension nDim of the correlation matrix. ； (6) Convert the Q and R array matrices of each pixel into one-dimensional arrays Mx and My of five elements, and concatenate and merge them to obtain the position matrix M; (7) Calculate the Gamma parameter table ； Gamma expansion and 0 filling are convenient for the next step: ； (8) Calculate the bias matrix Bias: Traverse each pixel of the image, calculate the bias parameter value Sum, and obtain the bias matrix Bias: ； (9) Bias correction: convert the original image to the logarithmic domain and then perform subtraction ； (10) Exponential transformation to obtain the corrected image: ； (11) Calculate the pot cover matrix template GGMask: ； (12) Divide the pot cover template into a distance ring RingImg with the center pixel of the image as the center according to the distance difference; (13) Calculate the pixel mean of each ring and generate the mean list array RingGrayMean of the distance to the center of the image; (14) Perform neighborhood smoothing correction on the distance list array RingGrayMean to generate a new mean list RingGrayMeanSmooth; (15) Distance-weighted interpolation is performed on the mean list RingGrayMeanSmooth, and the distance of each pixel in the image is included to obtain the final pot cover distance parameter table DistTable.

3. The method for removing pot lid from dynamic scene infrared image according to claim 1, characterized in that: The pot cover parameter table will be cleared when the power is off, and the pot cover acquisition instruction will be sent again, and the pot cover parameter table will be calculated once using the current scene image to replace the previous pot cover parameter table.

Images