Infrared small target rapid detection method based on gray scale descent angle contrast algorithm

Abstract

The invention belongs to the field of image processing calculation, and particularly relates to an infrared small target rapid detection method based on a gray scale descent angle contrast algorithm. The method comprises the following steps: S1, acquiring a to-be-detected image, setting a sliding window, and scanning the to-be-detected image by using the sliding window; s2, taking the product of the evaluation result # imgabs0 #, the difference calculation result # imgabs1 # and the noise reduction factor # imgabs2 # as a gray-scale descent angle contrast value; and S3, calculating a gray-scale descent angle contrast value corresponding to the sliding window of each scanning position, and taking an area in which the gray-scale descent angle contrast value is greater than zero as an area in which the infrared small target is located, thereby realizing rapid detection of the infrared small target. The method provided by the invention has the advantages that excellent detection capability and excellent real-time performance are both considered, and noise can be effectively resisted.

Description

Technical Field

[0001] The invention belongs to the field of image processing and calculation, and in particular relates to a method for quickly detecting small infrared targets based on a grayscale decreasing angle contrast algorithm. Background Art

[0002] The infrared small target detection method has the advantages of good real-time performance, strong detectability, good portability, the ability to detect radar blind spots, strong ability to penetrate smoke and dust, and the ability to work day and night. It also has the characteristics of high detection rate, low false alarm rate, and fast detection speed. It is widely used in civil and military fields, such as precision guidance, surveillance, and remote sensing, and can effectively detect and identify small targets such as drones.

[0003] In recent years, researchers have introduced the human visual system (HVS) into the study of infrared small target detection. HVS is a layered visual processing system composed of an optical system, retina, and visual perception pathways. HVS can quickly locate significant targets in complex external scenes. Although infrared weak targets are very small, they still have certain differences relative to the local background, and HVS can use these differences to quickly locate these areas. Therefore, the current mainstream HVS algorithms are mainly based on local contrast methods: Among them, the mean absolute gray difference (AAGD) algorithm uses the structure of two nested windows to calculate the contrast by calculating the difference between the average values ​​of the inner and outer windows. It has low computational complexity and is suitable for practical applications; the relative local contrast (RLCM) method considers the grayscale difference between the target and the adjacent background, and calculates the contrast by selecting the grayscale value sorting around each pixel to enhance the target or suppress the background; the three-layer template local difference measurement (TTLDM) method is adopted, and a three-layer template local difference measurement algorithm combining grayscale difference measurement and variance difference measurement is proposed, which realizes the target enhancement and background suppression of infrared weak targets; the infrared small target detection (ELUM) method based on local component uncertainty measurement and consistency evaluation performs target detection through uncertainty measurement of local components and signal enhancement based on energy weighted function. However, these methods still face great challenges in balancing excellent detection capabilities and good real-time performance for infrared small target detection in complex backgrounds. In addition, the impact of noise on detection performance is also a problem that cannot be ignored. Summary of the invention

[0004] In view of this, the present invention aims to provide a method for rapid detection of infrared small targets based on grayscale decreasing angle contrast algorithm, so as to solve the problem that the existing technology cannot balance excellent detection capability and excellent real-time performance for infrared small target detection under complex background. The present invention provides a method for balancing excellent detection capability and excellent real-time performance, and can effectively resist noise.

[0005] To achieve the above object, the technical solution created by the present invention is implemented as follows: A method for rapid detection of infrared small targets based on grayscale decreasing angle contrast algorithm specifically comprises the following steps: S1: Acquire the image to be detected, set a sliding window, and use the sliding window to scan the image to be detected; S2: Use the energy residual function to calculate the sliding window of each scanning position and pre-evaluate the target area of ​​each sliding window ; The minimum difference between the central pixel of the sliding window at each scanning position and the double-layer local neighborhood pixels is calculated by minimizing the difference function to obtain the difference calculation results of each sliding window. ; Calculate the double grayscale drop angle corresponding to the sliding window of each scanning position, and calculate the double grayscale drop angle tangent quotient corresponding to the sliding window of each scanning position based on the double grayscale drop angle, and obtain the noise reduction factor corresponding to the sliding window of each scanning position based on the double grayscale drop angle tangent quotient; The results of the evaluation , Difference calculation results and noise reduction factor The product of is taken as the grayscale drop angle contrast value; S3: Calculate the grayscale decreasing angle contrast value corresponding to the sliding window of each scanning position, and take the area where the grayscale decreasing angle contrast value is greater than zero as the area where the infrared small target is located, so as to achieve rapid detection of the infrared small target.

[0006] Furthermore, the images to be detected include the GOT-10k dataset.

[0007] Furthermore, in step 2, the energy residual of the current sliding window is calculated by the following formula to pre-evaluate the target area of ​​the current sliding window: ; ; ; in, is the grayscale matrix reconstructed with the average pixel value of the current sliding window, is a DOG filter, G is a two-dimensional Gaussian function with normal distribution, x and y are the horizontal and vertical coordinates of the two-dimensional coordinate system in the two-dimensional Gaussian function, and All are standard deviations.

[0008] Furthermore, in step S2, when the target size is n×n, the size of the inner local neighborhood pixels is (n+2)×(n+2), and the size of the outer local neighborhood pixels is (n+4)×(n+4).

[0009] Furthermore, in step 2, the minimum difference between the central pixel of the current sliding window and the double-layer local neighborhood pixels is calculated by the following formula: ; in, is the difference calculation result, is the gray value of the center pixel, is the image to be detected, is the inner local neighborhood pixel, is the outer local neighborhood pixel, and ⊕ is the dilation operation.

[0010] Furthermore, in step S2, the dual grayscale decreasing angle of the current sliding window includes a first-order neighborhood grayscale decreasing angle and a second-order neighborhood grayscale decreasing angle; a pixel set corresponding to the 8 neighborhood pixel positions adjacent to the center pixel of the current sliding window is defined as a first-order neighborhood, and any layer of local features in the current sliding window except the center pixel and the first-order neighborhood is taken as a second-order neighborhood, and the size of the second-order neighborhood is larger than the target size n×n.

[0011] Furthermore, in step S2, the noise reduction factor is calculated The steps include: Calculate the first-order neighborhood grayscale drop angle: ; in, is the first-order neighborhood grayscale drop angle, is the difference between the gray value of the central pixel in the current sliding window and the maximum gray value of the pixel in the first-order neighborhood. is the gray value of the center pixel in the current sliding window, is the maximum grayscale value of the pixel in the first-order neighborhood; Calculate the second-order neighborhood grayscale drop angle: ; in, is the second-order neighborhood grayscale drop angle, is the difference between the gray value of the central pixel in the current sliding window and the maximum gray value of the pixel in the second-order neighborhood. is the gray value of the center pixel in the current sliding window, is the maximum grayscale value of the pixel in the second-order neighborhood; Based on the first-order neighborhood grayscale drop angle and the second-order neighborhood grayscale drop angle, calculate the double grayscale drop angle tangent quotient of the current sliding window : , ; Based on double grayscale descending angle tangent quotient , calculate the noise reduction factor of the current sliding window: .

[0012] Furthermore, the size of the sliding window is larger than the target size.

[0013] Compared with the prior art, the invention can achieve the following beneficial effects: The present invention creates the infrared small target rapid detection method based on the grayscale descending angle contrast algorithm, calculates the energy residual of the image through the energy residual function, and performs two target positioning using the minimum difference function, thereby ensuring the credibility of the target information, distinguishing the target from the salt noise through the structural characteristics of the dual grayscale descending angle detection model, and further enhancing the target and suppressing the background, which can greatly ensure the accuracy of target detection. In addition, the dual grayscale descending angle tangent quotient is calculated in the layer ring area of ​​the first-order neighborhood and the second-order neighborhood, which significantly improves the operation efficiency, reduces the operation complexity, maintains excellent detection capability, and the present invention has the ability to resist salt noise interference. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The drawings constituting part of the present invention are used to provide a further understanding of the present invention. The exemplary embodiments and descriptions of the present invention are used to explain the present invention and do not constitute an improper limitation on the present invention. In the drawings: Figure 1 A schematic diagram of a process flow of a method for rapid detection of infrared small targets based on a grayscale decreasing angle contrast algorithm according to an embodiment of the present invention; Figure 2 A schematic diagram of the structure of the grayscale decreasing angle contrast algorithm described in the embodiment of the present invention; Figure 3 A schematic diagram of a model of the grayscale drop angle described in an embodiment of the present invention; Figure 4 A schematic diagram of a model of a double grayscale drop angle according to an embodiment of the present invention; Figure 5 A schematic diagram of a grayscale matrix slice structure of a target area under the premise that the target centroid is larger than the neighborhood of the target area as described in an embodiment of the present invention; Figure 6 A schematic diagram of a grayscale matrix slice structure of a target area under the premise that the target centroid is equal to the neighborhood of the target area as described in an embodiment of the present invention; Figure 7 A schematic diagram of a grayscale matrix slice structure of a noise region according to an embodiment of the present invention; Figure 8 A schematic diagram of a grayscale matrix slice structure of a target outer edge region according to an embodiment of the present invention; Fig. 9 A schematic diagram of the grayscale matrix slicing structure of the continuous strong edge area described in the embodiment of the present invention. DETAILED DESCRIPTION

[0015] In order to make the purpose, technical solution and advantages of the invention more clear, the invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described here are only used to explain the invention and do not constitute a limitation of the invention.

[0016] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0017] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In addition, the terms "first", "second", etc. are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, features defined as "first", "second", etc. may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "multiple" means two or more.

[0018] In the description of the invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installation", "connection" and "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the invention can be understood according to specific circumstances.

[0019] The present invention will be described in detail below with reference to the accompanying drawings and in combination with embodiments.

[0020] like Figure 1-Figure 2 As shown, the present invention proposes a method for rapid detection of infrared small targets based on grayscale decreasing angle contrast algorithm, including: S1: acquiring an image to be detected, setting a sliding window, and scanning the image to be detected using the sliding window; S2: calculating the sliding window at each scanning position using an energy residual function, and pre-evaluating the target area of ​​each sliding window ; The minimum difference between the central pixel of the sliding window at each scanning position and the double-layer local neighborhood pixel is calculated by minimizing the difference function to obtain the difference calculation results of each sliding window ; Calculate the double grayscale drop angle corresponding to the sliding window of each scanning position, and calculate the double grayscale drop angle tangent quotient corresponding to the sliding window of each scanning position based on the double grayscale drop angle, and obtain the noise reduction factor corresponding to the sliding window of each scanning position based on the double grayscale drop angle tangent quotient; The evaluation result , Difference calculation results and noise reduction factor The product of is taken as the grayscale decrease angle contrast value; S3: Calculate the grayscale decrease angle contrast value corresponding to the sliding window of each scanning position, and take the area where the grayscale decrease angle contrast value is greater than zero as the area where the infrared small target is located, so as to realize the rapid detection of the infrared small target.

[0021] It should be noted that the present invention proposes a dual grayscale decreasing angle contrast algorithm (DGDACM) to simultaneously enhance the target and adaptively suppress complex clutter and salt noise. DGDACM is mainly divided into three parts. First, based on the inconsistency between the target and the background in the local grayscale neighborhood, the present invention calculates the energy residual of the original image through the energy residual function to pre-evaluate the potential target area; secondly, based on the grayscale characteristics of the local neighborhood, the minimum difference between the central pixel and the local neighborhood is calculated through the minimum difference function, and the level difference between the central pixel and the neighborhood is calculated to enhance and re-locate the potential target area, and suppress the background area; finally, based on the dual grayscale decreasing angle anti-noise detection model, the target is further enhanced, and the flat background area, the neighborhood outside the target, the continuous strong edge and the salt noise area are further suppressed. Finally, adaptive threshold segmentation is used to obtain accurate target detection results.

[0022] In some instances, the image to be detected generally includes a GOT-10k dataset.

[0023] It should be noted that, in practical applications, the image to be detected is calculated by defining templates of multiple sizes, and finally the optimal solution is taken to obtain the infrared weak target.

[0024] In some examples, in step 2, the energy residual of the current sliding window is calculated by the following formula to pre-evaluate the target area of ​​the current sliding window: ; ; ; in, is the grayscale matrix reconstructed with the average pixel value of the current sliding window, is a DOG filter, G is a two-dimensional Gaussian function with normal distribution, x and y are the horizontal and vertical coordinates of the two-dimensional coordinate system in the two-dimensional Gaussian function, and All are standard deviations.

[0025] It should be noted that based on the radiation characteristics of small infrared targets, in the local neighborhood of the infrared target, the grayscale value of the central pixel is the largest, and the grayscale change between the central pixel and the central neighborhood is not obvious. The grayscale of the inner and outer edges of the small infrared target changes dramatically. A preliminary assessment of potential target areas can be made.

[0026] In some instances, in step S2, when the target size is n×n, the inner local neighborhood pixels ( Figure 1 The size of the inner layer is (n+2)×(n+2), and the outer local neighborhood pixels ( Figure 1 The size of the outer structure (abbreviated as the outer structure) is (n+4)×(n+4).

[0027] It should be noted that, based on the grayscale characteristics of small infrared targets, the target centroid is the maximum grayscale value in the local neighborhood, which must be greater than the grayscale values ​​of all pixels in the local neighborhood. Therefore, the present invention can enhance the target and suppress the background by calculating the difference between each pixel point and the maximum pixel between different layers.

[0028] In some examples, in step 2, the minimum difference between the central pixel of the current sliding window and the pixels of the double-layer local neighborhood is calculated by the following formula: ; in, is the difference calculation result, is the gray value of the center pixel, is the image to be detected, is the inner local neighborhood pixel, is the outer local neighborhood pixel, and ⊕ is the dilation operation.

[0029] It should be noted that The result is the maximum value of the pixels in the inner local neighborhood. The result is the pixel maximum value of the outer local neighborhood pixels.

[0030] Furthermore, H(x) is a step function, defined as follows: ; By calculation , so that the target area is enhanced because the difference between the pixel center and the maximum pixel between different layers is positive, and other areas such as the background are suppressed because the difference between the pixel center and the maximum pixel between different layers is negative. Therefore, It can effectively enhance the target while suppressing the background, and can perform secondary positioning of the target area and eliminate most of the background interference.

[0031] In some instances, in step S2, the dual grayscale decreasing angle of the current sliding window includes a first-order neighborhood grayscale decreasing angle and a second-order neighborhood grayscale decreasing angle; a set of pixels corresponding to the 8 neighborhood pixel positions adjacent to the center pixel of the current sliding window is defined as a first-order neighborhood, and any layer of local features in the current sliding window except the center pixel and the first-order neighborhood is taken as a second-order neighborhood, and the size of the second-order neighborhood is slightly larger than the target size n×n.

[0032] It should be noted that the pixel set corresponding to the 8 neighboring pixel positions adjacent to the central pixel is defined as the first layer of local features, the pixel set corresponding to the 16 neighboring pixel positions adjacent to the 8 neighboring pixel positions is defined as the second layer of local features, ..., the principle of obtaining any layer of local features here is the same as above.

[0033] Furthermore, based on the relationship between the central pixel of the current sliding window and the first-order neighborhood and the second-order neighborhood, the first-order neighborhood grayscale drop angle and the second-order neighborhood grayscale drop angle are calculated. Here, the grayscale drop angle is explained as follows: Figure 3 As shown in the 3D grayscale image, the grayscale direction of the pixel at the center is a vertical edge (such as Figure 3 The red dotted line in the middle shows that the line connecting the pixel at the center and the pixel corresponding to the maximum grayscale value in the first-order neighborhood is a hypotenuse (as shown in Figure 3 The angle formed by the green dashed line in the middle) (non-obtuse angle, such as Figure 3 As shown by the purple arc in the figure), it is defined as the grayscale decreasing angle, and when the angle exceeds 𝝅 / 2, there will be no decreasing trend. Therefore, in the definition, the grayscale of pixels at the positions where these angles that do not meet the definition are located is reset to 0.

[0034] In some examples, the noise reduction factor is calculated The steps include: Calculate the first-order neighborhood grayscale drop angle: ; in, is the first-order neighborhood grayscale drop angle, is the difference between the gray value of the central pixel in the current sliding window and the maximum gray value of the pixel in the first-order neighborhood. is the gray value of the center pixel in the current sliding window, is the maximum grayscale value of the pixel in the first-order neighborhood; Calculate the second-order neighborhood grayscale drop angle: ; in, is the second-order neighborhood grayscale drop angle, is the difference between the gray value of the central pixel in the current sliding window and the maximum gray value of the pixel in the second-order neighborhood. is the gray value of the center pixel in the current sliding window, is the maximum grayscale value of the pixel in the second-order neighborhood; Based on the first-order neighborhood grayscale drop angle and the second-order neighborhood grayscale drop angle, calculate the double grayscale drop angle tangent quotient of the current sliding window : , ; Based on double grayscale descending angle tangent quotient , calculate the noise reduction factor of the current sliding window: .

[0035] It should be noted that the calculation method of the double grayscale drop angle is as follows Figure 4 As shown in , the red area represents the center of the current detection position, the blue area represents the first-order neighborhood, and the yellow area represents the second-order neighborhood. Figure 4 As shown in the figure, the red straight line represents the first-order vertical edge and the second-order vertical edge, the green straight line represents the first-order oblique edge and the second-order oblique edge, the blue straight line represents the maximum grayscale of the first-order neighborhood, the yellow straight line represents the maximum grayscale of the second-order neighborhood, and the purple arc represents the grayscale drop angle of the first-order neighborhood and the grayscale drop angle of the second-order neighborhood. Furthermore, regarding the selection of the template size: the template size of the first-order neighborhood is fixed to 3×3, while the template size of the second-order neighborhood only needs to be equal to or slightly larger than the target size.

[0036] In some embodiments, the size of the sliding window is larger than the target size.

[0037] It should be noted that in the calculation of step S2, the evaluation result is calculated , Difference calculation results The size of the sliding window used by the noise reduction factor DGDγ is usually the same. When the size of the sliding window is different, the method of the present invention can also achieve rapid detection of infrared small targets. The sliding window size is determined by the center window. The center window of the optimal size must wrap the target signal. The complete minimum window structure has a center window size of 3×3, and the size is adjustable. Adjusting the size of the sliding window can effectively process targets of different sizes and shapes.

[0038] According to the infrared radiation characteristics of small infrared targets, the significance of the target is detected, and image areas such as sand, stone, soil or hills are classified as backgrounds. The brightness of these backgrounds depends on their ability to radiate infrared waves and is related to their materials. In general, the background is usually a dark area. Some backgrounds with strong infrared wave radiation capabilities appear as brighter areas, but unlike the grayscale features of small targets highlighted in local neighborhoods (greater than 1 pixel), the backgrounds in brighter areas are generally continuous over a large range. In addition, the grayscale features of salt noise are only highlighted within 1 pixel. Therefore, the present invention takes real local grayscale matrix slices as an example, conducts a qualitative analysis of the detection of the DGDACM algorithm in different areas, discusses the performance of the DGDACM algorithm in different detection areas, and illustrates the effectiveness of the DGDACM algorithm.

[0039] Because I ER Similar to the Gaussian difference filter, it can preliminarily screen the boundaries of local areas, determine the possible targets in the local areas, and then retain and enhance these boundary areas. Therefore, it will enhance the target area, the noise area, the outer edge of the target area and the outer edge of the continuous highlight area, and suppress the continuous highlight inner area and the slowly changing background area.

[0040] (1) Target area (the target centroid is larger than the centroid neighborhood within the target area) When the gray value of the target centroid is greater than some gray values ​​in the target area (i.e., the centroid is the only gray value maximum), such as Figure 5 As shown, since the target area has a boundary, I ER Be enhanced.

[0041] ; In the target area, I MD Also enhanced.

[0042] = = ; At this time, ∆h1>0, the first-order neighborhood grayscale drop angle In the interval (0, π / 4), the tangent value is 1 / 16, which is a constant greater than 0.

[0043] = = ; The second-order descent angle Also in the interval (0, π / 4), the tangent value is 1 / 158, which is a constant greater than 0.

[0044] = ; The tangent quotient of the double grayscale drop angle is greater than 1. DGDγ is also greater than 1. Therefore, DGDACM effectively enhances the target area.

[0045] (2) Target area (the target centroid is equal to the centroid neighborhood within the target area) When the target centroid grayscale value is equal to some grayscale value in the target area (i.e., the centroid is not the only grayscale maximum value), Figure 6 As shown, there is a boundary in the target area, so the IER is enhanced.

[0046] ; In the target area, I MD It has also been enhanced; = = ; At this time, ∆h1=0, the first-order neighborhood grayscale drop angle In the interval (0, π / 2), the tangent value tends to +∞.

[0047] = = ; The second-order descent angle In the interval (0, π / 4), the tangent value is 1 / (158), which is a constant greater than 0.

[0048] = ; The double grayscale descent angle tangent quotient is much greater than 1. DGDγ is enhanced. Therefore, the DGDACM algorithm effectively enhances the target area.

[0049] (3) Noise area When the detection center is exactly at the noise position, such as Figure 7 As shown, there is a boundary in the noise region, so I ER Be enhanced.

[0050] ; In the noisy area, I MD Also enhanced.

[0051] = = ; At this time, ∆h1>0, the first-order grayscale drop angle In the interval (0, π / 4), the tangent value is 1 / 138, which is a constant greater than 0.

[0052] = = ; The second-order descent angle Also in the interval (0, π / 4), the tangent value is 1 / 128, which is a constant greater than 0.

[0053] = ; Due to the grayscale characteristics of the noise, the drop angles in the two-order neighborhood are similar, and the double grayscale drop angle tangent quotient is approximately equal to 1. In the noise area, although I ER and I MD The noise is enhanced, but DGDγ is approximately 0 and DGDACM is also approximately 0. Therefore, DGDACM effectively suppresses the noise region.

[0054] (4) Outer edge of the target area In the local grayscale matrix of the outer edge of the planar target, Figure 8 As shown, the white dotted area is the center position of the current pixel, the white solid line of the first-layer neighborhood is the first-order neighborhood, and the white solid line of the third-layer neighborhood is the second-order neighborhood. At this time, ∆h1<0, and the grayscale drop angle of the first-order neighborhood is an obtuse angle. According to the definition, the double grayscale drop angle tangent quotient (γ) is equal to 0. There is a boundary at the outer edge of the target area, so I ER Be enhanced.

[0055] IMD=H[(74-255)]·H((74-151))=0(22); At the outer edge of the target area, IMD is suppressed, ∆h1<0, the first-order neighborhood gray value drop angle is obtuse, and by definition, the tangent quotient (γ) of the second-order neighborhood gray value drop angle is equal to 0, DGDγ is equal to 0, and DGDACM is also equal to 0. Therefore, DGDACM effectively suppresses the outer edge of the target area.

[0056] (5) Continuous strong edge regions In the local grayscale matrix of a plane continuous strong edge, when ∆h1<0, according to the definition, the double grayscale descent angle tangent quotient (γ) is equal to 0, so here we only need to discuss the situation of ∆h1≥0. According to the grayscale characteristics of continuous strong edges, in the local area, the grayscale change is not obvious, and the grayscale difference is approximately 0, so it can be considered that ∆h1≈0. Therefore, here we take the situation when ∆h1=0 as the grayscale matrix ( Fig. 9 ) for discussion, such as Fig. 9 shown.

[0057] There is a boundary in the continuous strong edge region, but the detection center is in the continuous highlight edge region, so I ER is suppressed.

[0058] ; In the continuous strong edge region, IMD is suppressed.

[0059] = = ; In the detection area of ​​continuous strong edges, since ∆h1=0, the first-order neighborhood grayscale decrease angle tends to π / 2+, and the tangent value tends to positive infinity.

[0060] = = ; At the same time, ∆h2=0, the second-order grayscale decrease angle also tends to π / 2+, and the tangent value tends to positive infinity.

[0061] = ; The tangent quotient at this time When x approaches π / 2+, it is equivalent to infinity, so γ = 1, DGDγ is equal to 0, and DGDACM is also equal to 0. Therefore, DGDACM effectively suppresses the continuous strong edge region.

[0062] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the disclosure of the present invention can be performed in parallel, sequentially or in different orders, as long as the desired results of the technical solution disclosed in the present invention can be achieved, and this document does not limit this.

[0063] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for rapid detection of infrared small targets based on grayscale decreasing angle contrast algorithm, characterized by: The specific steps include: S1: Acquire the image to be detected, set a sliding window, and use the sliding window to scan the image to be detected; S2: Use the energy residual function to calculate the sliding window of each scanning position and pre-evaluate the target area of ​​each sliding window ; The minimum difference between the central pixel of the sliding window at each scanning position and the double-layer local neighborhood pixels is calculated by minimizing the difference function to obtain the difference calculation results of each sliding window. ; Calculate the double grayscale drop angle corresponding to the sliding window of each scanning position, and calculate the double grayscale drop angle tangent quotient corresponding to the sliding window of each scanning position based on the double grayscale drop angle, and obtain the noise reduction factor corresponding to the sliding window of each scanning position based on the double grayscale drop angle tangent quotient; The results of the evaluation , Difference calculation results and noise reduction factor The product of is taken as the grayscale drop angle contrast value; S3: Calculate the grayscale decreasing angle contrast value corresponding to the sliding window of each scanning position, and take the area where the grayscale decreasing angle contrast value is greater than zero as the area where the infrared small target is located, so as to achieve rapid detection of the infrared small target.

2. The infrared small target rapid detection method based on grayscale decreasing angle contrast algorithm according to claim 1 is characterized in that: The image to be detected includes a GOT-10k data set.

3. The infrared small target rapid detection method based on grayscale decreasing angle contrast algorithm according to claim 1 is characterized in that: In step 2, the energy residual of the current sliding window is calculated by the following formula to pre-evaluate the target area of ​​the current sliding window: ； ； ； in, is the grayscale matrix reconstructed with the average pixel value of the current sliding window, is a DOG filter, G is a two-dimensional Gaussian function with normal distribution, x and y are the horizontal and vertical coordinates of the two-dimensional coordinate system in the two-dimensional Gaussian function, and All are standard deviations.

4. The infrared small target rapid detection method based on grayscale decreasing angle contrast algorithm according to claim 1 is characterized in that: In step S2, when the target size is n×n, the size of the inner local neighborhood pixels is (n+2)×(n+2), and the size of the outer local neighborhood pixels is (n+4)×(n+4).

5. The infrared small target rapid detection method based on grayscale decreasing angle contrast algorithm according to claim 4 is characterized in that: In step 2, the minimum difference between the central pixel of the current sliding window and the double-layer local neighborhood pixels is calculated by the following formula: ； in, is the difference calculation result, is the gray value of the center pixel, is the image to be detected, is the inner local neighborhood pixel, is the outer local neighborhood pixel, and ⊕ is the dilation operation.

6. The infrared small target rapid detection method based on grayscale decreasing angle contrast algorithm according to claim 1 is characterized by: In step S2, the dual grayscale decreasing angle of the current sliding window includes a first-order neighborhood grayscale decreasing angle and a second-order neighborhood grayscale decreasing angle; a pixel set corresponding to the 8 neighborhood pixel positions adjacent to the center pixel of the current sliding window is defined as a first-order neighborhood, and any layer of local features in the current sliding window except the center pixel and the first-order neighborhood is used as a second-order neighborhood, and the size of the second-order neighborhood is larger than the target size n×n.

7. The infrared small target rapid detection method based on grayscale decreasing angle contrast algorithm according to claim 6 is characterized in that: In step S2, the noise reduction factor is calculated The steps include: Calculate the first-order neighborhood grayscale drop angle: ； in, is the first-order neighborhood grayscale drop angle, is the difference between the gray value of the central pixel in the current sliding window and the maximum gray value of the pixel in the first-order neighborhood. is the gray value of the center pixel in the current sliding window, is the maximum grayscale value of the pixel in the first-order neighborhood; Calculate the second-order neighborhood grayscale drop angle: ； in, is the second-order neighborhood grayscale drop angle, is the difference between the gray value of the central pixel in the current sliding window and the maximum gray value of the pixel in the second-order neighborhood. is the gray value of the center pixel in the current sliding window, is the maximum grayscale value of the pixel in the second-order neighborhood; Based on the first-order neighborhood grayscale drop angle and the second-order neighborhood grayscale drop angle, calculate the double grayscale drop angle tangent quotient of the current sliding window : , ； Based on double grayscale descending angle tangent quotient , calculate the noise reduction factor of the current sliding window: 。 8. The infrared small target rapid detection method based on grayscale decreasing angle contrast algorithm according to claim 1 is characterized by: The size of the sliding window is larger than the target size.

Images