An adaptive bias-updating infrared image scene correction method

Abstract

The present invention relates to an adaptive offset-updating infrared image scene correction method, belonging to the field of infrared image processing technology. The present invention uses a sequence of 14-bit raw infrared image frames to perform pixel-level motion estimation on the image frame sequence, updates the offset correction matrix for image frames determined to be in motion, and corrects the image frame sequence using the offset correction matrix to reduce image non-uniformity. While removing image non-uniformity, the present invention can better preserve image detail without introducing image ghosting, has low algorithm complexity, and can be implemented in an embedded system.

Description

Technical Field

[0001] The invention belongs to the technical field of infrared image processing, and in particular relates to an adaptive bias-updating infrared image scene correction method. Background Art

[0002] Infrared thermal imagers can exhibit non-uniformity in infrared images due to factors such as the semiconductor manufacturing process of the detector. Domestic detectors are particularly prone to exhibiting significant non-uniformity when detecting weak radiation scenes due to poor detector responsivity. Furthermore, after prolonged operation, each pixel in the detector will experience varying degrees of temperature drift, which can also introduce non-uniformity. Therefore, addressing this non-uniformity issue is crucial to ensure image quality.

[0003] Currently, methods for correcting non-uniformity in infrared images can be divided into two categories: reference-source-based NUC correction and scene-based NUC correction. Reference-source-based NUC correction methods include baffle correction, blackbody correction, and lag correction. These methods work by placing a uniformly radiating object in front of the detector or thermal imager to block it, causing the detector pixels to exhibit the same response to objects radiating at the same temperature, resulting in better image uniformity. However, these methods can interrupt the normal operation of the thermal imager, easily affecting tasks such as target search, observation, and tracking during use.

[0004] The scene-based NUC correction method uses image algorithms to correct image non-uniformity, and there will be no interruption in the imaging process of the thermal imager. The current mainstream methods can be summarized into three categories: time-domain high-pass filtering, constant statistical method, and neural network method. After reading the literature and simulating the algorithm, it is found that the ghosting of the time-domain high-pass filtering method is difficult to eliminate, and weak details will be lost; the neural network method has high algorithm complexity, is not easy to implement in engineering, and is complex to adjust the parameters of the adapted scene; the constant statistical method is based on the assumption that the response of adjacent pixels observing the same radiation scene should be the same, and uses deviation correction. The algorithm complexity is low and can be implemented in an embedded manner, but the algorithm convergence and ghosting problems need to be solved. At present, there is no appropriate scene correction method that can solve the above problems. Summary of the Invention

[0005] In order to solve the problems of difficulty in eliminating ghost images, loss of image details, high algorithm complexity, etc. in the above-mentioned scene-based NUC correction method, the present invention proposes an adaptive bias update infrared image scene correction method.

[0006] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0007] An adaptive bias-updating infrared image scene correction method comprises the following steps:

[0008] Step 1: Input infrared 14-bit raw image frame sequence n represents the nth frame, and the current frame image is The previous frame image is

[0009] Step 2: For the current frame image input in step 1, use the correction matrix C calculated in the previous frame n-1 Perform pre-correction of the current frame, pre-correction of the current frame image

[0010] Step 3: Pre-correct the image of the current frame Correct the image with the previous frame The motion threshold T and frame difference image D are calculated using the inter-frame motion adaptive parameter estimation method;

[0011] Step 4: Based on the motion threshold T and frame difference image D obtained in step 3, perform pixel-level motion detection estimation and calculate the pixel-level update area At the same time, the Euler distance is used to determine the motion of the frame difference image D boundary to determine whether the image scene is moving;

[0012] Step 5: For the image determined as a motion frame in step 4, use the offset update algorithm to correct the offset matrix C n Update; for the image determined as a non-motion frame in step 4, the bias correction matrix C n Keep with C n-1 same;

[0013] Step 6: Use the updated region obtained in step 4 And the bias correction matrix C obtained in step 5 n Correct the current image frame to obtain the corrected image frame The correction method is

[0014] Step 7: After performing image algorithm enhancement on the corrected image frame obtained in step 6, the histogram mapping method is used to convert the 14-bit corrected image frame into an 8-bit image that can be observed by the human eye.

[0015] Furthermore, preferably, the inter-frame motion adaptive parameter estimation method adopted in step 3 is used to calculate the motion threshold T using a 14-bit infrared image signal-to-noise ratio statistical method, and the calculation method is: The constant is the signal-to-noise ratio of the infrared image; the frame difference image D is the pre-corrected image of the current frame. Correct the image with the previous frame The difference in data, the frame difference image is

[0016] Further, preferably, in step four, the pixel value D of the frame difference image D (i,j) is compared with a threshold value T. When D (i,j) > T, the image pixel is determined to be moving and is marked as 1; when D (i,j) < T, the image pixel is determined to be not moving and is marked as 0, obtaining a pixel-level updated area where (i, j) is the pixel position; at the same time, the Euler distance calculation method is used for the frame difference image D to determine whether the image scene is moving. The calculation method is where D up , D down , D left , D right are the pixel values of the first row, the last row, the first column, and the last column of the frame difference image respectively. When d > C, the image frame is determined to be a moving frame; when d < C, the image frame is determined to be a non-moving frame, and C is a constant.

[0017] Further, preferably, the offset correction matrix C n adopted in step five is updated by an algorithm, and its calculation method is where is the offset correction matrix after updating for the current frame, is the offset correction matrix of the previous frame, and diff(i, j) is the deviation value between the pixel value to be updated and the neighborhood pixel value where (i, j) is the pixel position and norm is a constant.

[0018] Further, preferably, the image enhancement algorithm adopted in step seven includes gamma correction, image sharpening, digital detail enhancement, and histogram mapping algorithms.

[0019] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0020] Compared with the existing infrared image non-uniformity correction technology, the present invention can directly use the image frame sequence to estimate whether the image is moving, without adding a motion detection sensor, can well eliminate the ghost problem brought by the algorithm iteration of the static frame, and the processing process of the present invention is more concise, with a small amount of calculation and is convenient for hardware implementation. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, form a part of this application, and do not constitute a limitation on the embodiments of the present invention. In the drawings:

[0022] Figure 1 is a schematic flow chart of the method of the present invention.

[0023] Figure 2 is the infrared input image that has not been corrected in the present invention;

[0024] Figure 3 The infrared output image after correction processing by the present invention;

[0025] Figure 4 This is the non-uniform part of the image removed by the present invention. DETAILED DESCRIPTION

[0026] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with examples and drawings. The exemplary embodiments of the present invention and their descriptions are only used to explain the present invention and are not intended to limit the present invention.

[0027] Those skilled in the art will understand that the following examples are intended to illustrate the present invention only and should not be construed as limiting the scope of the present invention. Where specific techniques or conditions are not specified in the examples, the techniques or conditions described in the literature in the art or in the product specifications were used. Materials or equipment used without manufacturer identification are commercially available conventional products.

[0028] like Figure 1 As shown, a method for adaptive bias updating infrared image scene correction includes the following steps:

[0029] Step 1: Input infrared 14-bit raw image frame sequence n represents the nth frame, and the current frame image is The previous frame image is

[0030] Step 2: For the current frame image input in step 1, use the correction matrix C calculated in the previous frame n-1 Perform pre-correction of the current frame, pre-correction of the current frame image

[0031] Step 3: Pre-correct the image of the current frame Correct the image with the previous frame The motion threshold T and frame difference image D are calculated using the inter-frame motion adaptive parameter estimation method;

[0032] Step 4: Based on the motion threshold T and frame difference image D obtained in step 3, perform pixel-level motion detection estimation and calculate the pixel-level update area At the same time, the Euler distance is used to determine the motion of the frame difference image D boundary to determine whether the image scene is moving;

[0033] Step 5: For the image determined as a motion frame in step 4, use the offset update algorithm to correct the offset matrix Cn Update; for the image determined as a non-moving frame in step four, the offset correction matrix C n Remains the same as C n-1 ;

[0034] Step six, use the updated area obtained in step four and the offset correction matrix C obtained in step five n to correct the current image frame and obtain a corrected image frame The correction method is

[0035] Step seven, after enhancing the corrected image frame obtained in step six using an image algorithm, use the histogram mapping method to convert the 14-bit corrected image frame into an 8-bit image observable by the human eye

[0036] Among them, for the inter-frame motion adaptive parameter estimation method adopted in step three, the motion threshold T is calculated using the 14-bit infrared image signal-to-noise ratio statistical method, and its calculation method is where the constant is obtained by statistically calculating the signal-to-noise ratio of the infrared image; at the same time, the frame difference image D is the pre-corrected image of the current frame calculated and the corrected image of the previous frame The difference in data, and the frame difference image is

[0037] In step four, the pixel value D of the frame difference image D (i,j) is compared with the threshold T. When D (i,j) > T, the image pixel is determined to be moving and marked as 1; when D (i,j) < T, the image pixel is determined to be non-moving and marked as 0, obtaining a pixel-level updated area where (i, j) is the pixel position; at the same time, use the Euler distance calculation method for the frame difference image D to determine whether the image scene is moving, and its calculation method is where D up 、D down 、D left 、D right are the pixel values of the first row, last row, first column, and last column of the frame difference image respectively. When d > C, the image frame is determined to be a moving frame; when d < C, the image frame is determined to be a non-moving frame, and C is a constant, preferably taking a value of 20

[0038] The offset correction matrix C adopted in step five n update algorithm, and its calculation method is norm, where is the updated offset correction matrix of the current frame, is the offset correction matrix of the previous frame, and diff(i, j) is the statistical pixel value to be updated and neighborhood pixel values Deviation value Where (i, j) is the pixel position and norm is a constant, preferably 100.

[0039] The image enhancement algorithms used in step seven include gamma correction, image sharpening, digital detail enhancement and histogram mapping algorithms.

[0040] The uncorrected infrared input image is as follows Figure 2 As shown, the image has severe non-uniformity, strong vertical stripes, and unclear scene details; Figure 3 The infrared output image after correction processing by the method of the present invention is obtained from Figure 3 It can be seen that the image non-uniformity is effectively removed and the scene information is clear; Figure 4 The image non-uniformity portion removed by the present invention is obtained by subtracting the input image from the output image, and is the image non-uniform portion removed using the method of the present invention.

[0041] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0042] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0043] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1The function specified in one or more boxes.

[0044] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0045] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An adaptive bias update infrared image scene correction method, characterized in that: The method comprises the following steps: Step 1: Input infrared 14-bit raw image frame sequence n represents the nth frame, and the current frame image is The previous frame image is Step 2: For the current frame image input in step 1, use the correction matrix C calculated in the previous frame n-1 Perform pre-correction of the current frame, pre-correction of the current frame image Step 3: Pre-correct the image of the current frame Correct the image with the previous frame The motion threshold T and frame difference image D are calculated using the inter-frame motion adaptive parameter estimation method; Step 4: Based on the motion threshold T and frame difference image D obtained in step 3, perform pixel-level motion detection estimation and calculate the pixel-level update area At the same time, the Euler distance is used to determine the motion of the frame difference image D boundary to determine whether the image scene is moving; Step 5: For the image determined as a motion frame in step 4, use the offset update algorithm to correct the offset matrix C n Update; for the image determined as a non-motion frame in step 4, the bias correction matrix C n Keep with C n-1 same; Step 6: Use the updated region obtained in step 4 And the bias correction matrix C obtained in step 5 n Correct the current image frame to obtain the corrected image frame The correction method is Step 7: After performing image algorithm enhancement on the corrected image frame obtained in step 6, the histogram mapping method is used to convert the 14-bit corrected image frame into an 8-bit image that can be observed by the human eye. In step 4, the pixel value D of the frame difference image D is (i，j) Compared with the threshold T, when D (i，j) >T, the image pixel is judged to be in motion and marked as 1; when D (i，j) <T, the image pixel is judged as not moving, marked as 0, and the pixel-level update area is obtained Where (i, j) is the pixel position; at the same time, the Euler distance calculation method is used on the frame difference image D to determine whether the image scene is moving. The calculation method is: Among them D up 、D down 、D left 、D right are the pixel values ​​of the first row, last row, first column, and last column of the frame difference image, respectively. When d>C, the image frame is determined to be a motion frame; when d<C, the image frame is determined to be a non-motion frame. C is a constant. The bias correction matrix C used in step 5 is n The update algorithm is calculated as follows: in is the updated bias correction matrix for the current frame, is the bias correction matrix of the previous frame, diff(i, j) is the statistical pixel value to be updated and neighborhood pixel values Deviation value Where (i, j) is the pixel position and norm is a constant.

2. The adaptive bias-updating infrared image scene correction method according to claim 1, characterized in that: The inter-frame motion adaptive parameter estimation method used in step 3 is used to calculate the motion threshold T using the 14-bit infrared image signal-to-noise ratio statistical method, which is calculated as follows: At the same time, the frame difference image D is used to calculate the pre-corrected image of the current frame. Correct the image with the previous frame The difference in data, the frame difference image is 3. The adaptive bias updating infrared image scene correction method according to claim 1, characterized in that: The image enhancement algorithms used in step seven include gamma correction, image sharpening, digital detail enhancement and histogram mapping algorithms.

Images