A method for detecting weak targets based on speed gear search

Abstract

This application relates to a weak target detection method based on speed-level search, comprising: acquiring a candidate target point set and a target search speed level set of a current batch of binarized image sequences, wherein the target search speed level set is calculated based on the maximum motion speed of the target to be detected and the target search step; for each candidate target point, searching within the corresponding search window of each frame of binarized image excluding the last frame at each target search speed level, and calculating the cumulative number of image frames searched for each candidate target point at each target search speed level; selecting the maximum cumulative number of image frames from the multiple cumulative number of image frames for each candidate target point and comparing it with a preset frame count decision threshold. Compared with multi-frame image-based methods that require a precise target motion model, this method significantly reduces its dependence on prior knowledge and is more applicable.

Description

Technical Field

[0001] This application relates to the field of infrared image processing technology, and in particular to a method for detecting weak targets based on speed level search. Background Technology

[0002] Infrared image target motion detection and tracking technology is a key technology for precision-guided weapon systems, space-based early warning systems, and target surveillance systems, playing a vital role in both military and civilian fields. The challenges of detecting weak infrared targets lie in: 1) Weak targets. At relatively long detection distances, targets with weak signal strength have limited distinguishability from the background and clutter on the detector; 2) Small targets. Small targets are typically considered point targets, where information such as appearance, size, and texture features are no longer available.

[0003] Since its inception, infrared target detection has developed into two main approaches. One is based on single-frame image algorithms, which are computationally simple and offer good real-time performance. However, for weak infrared targets, the small size of the target makes it difficult to distinguish it from background noise using only a single frame, limiting its practical application. Conversely, the other approach is based on multi-frame image sequences. These methods utilize target motion information and offer better detection performance under low signal-to-noise ratio conditions compared to single-frame methods, although their real-time performance is relatively poor. The core of image sequence-based methods is to accumulate target energy using target motion information, amplifying the difference between the target and the background, thus improving target detection. Representative methods include dynamic programming, 3D matched filtering, and recursive Bayesian methods. Most of these methods require accurate estimation of target motion information, which is often challenging. Inaccurate target motion models lead to decreased target detection performance. Summary of the Invention

[0004] Therefore, it is necessary to provide a weak target detection method based on speed gear search to address the above-mentioned technical problems. This method does not require a high degree of accuracy in estimating the target's motion information, while ensuring good target detection performance.

[0005] A method for detecting weak targets based on speed gear search, characterized in that the method includes:

[0006] Obtain a candidate target point set for the current batch of binarized image sequences; the candidate target point set is obtained by selecting pixels that have passed the binarization segmentation threshold in the last frame of the binarized image sequence.

[0007] Obtain a set of target search speed levels; the set of target search speed levels includes multiple target search speed levels; the set of target search speed levels is calculated based on the maximum movement speed of the target to be detected and the target search step; the movement speed of the target to be detected is in pixels per frame;

[0008] For each candidate target point in the candidate target point set, a search is performed in the corresponding search window of each frame of binarized image excluding the last frame binarized image at each target search speed level. The cumulative number of image frames searched for each candidate target point at each target search speed level is counted. The cumulative number of image frames is the total number of binarized image frames in the search window that contain pixels that have passed the binarization segmentation threshold.

[0009] The maximum cumulative image frame count is selected from the multiple cumulative image frame counts of each candidate target point, and the corresponding candidate target point is determined to be a target point based on whether the maximum cumulative image frame count meets a preset condition.

[0010] The aforementioned method for detecting weak targets based on speed level search includes: acquiring a candidate target point set of the current batch of binarized image sequences, wherein the candidate target point set is obtained by selecting pixels that have passed the binarization segmentation threshold in the last frame of the binarized image sequence; acquiring a target search speed level set, wherein the target search speed level set includes multiple target search speed levels; the target search speed level set is calculated based on the maximum motion speed of the target to be detected and the target search step; the motion speed of the target to be detected is in pixels / frame; and processing the candidate target point set... Each candidate target point is searched within its corresponding search window in each frame of the binary image (excluding the last frame) at its respective target search speed level. The cumulative number of image frames searched for each candidate target point at each target search speed level is counted. The cumulative number of image frames is the total number of binary images containing pixels that pass the binarization segmentation threshold within the search window. The maximum cumulative number of image frames is selected from the multiple cumulative number of image frames for each candidate target point. Whether the maximum cumulative number of image frames meets the preset conditions determines whether the corresponding candidate target point is a target point. For targets moving at a constant linear velocity, it is not necessary to know the target's accurate true velocity in advance. Only its maximum velocity is used as a constraint. An appropriate target search speed level is set, and each candidate target point is searched and counted within its corresponding search window at each target search speed level. Compared with methods based on multiple frames that require a precise target motion model, this method greatly reduces its dependence on prior knowledge and is more applicable. Attached Figure Description

[0011] Figure 1This is a flowchart illustrating a method for detecting weak targets based on speed gear search in one embodiment.

[0012] Figure 2 This is a schematic diagram of the search window corresponding to each frame of the binarized image when the target search level is (1.8,2] in one embodiment;

[0013] Figure 3 This is a schematic diagram of image batch processing when the sliding frame number is 1 frame in one embodiment;

[0014] Figure 4 This is a structural block diagram of a weak target detection device based on speed gear search in one embodiment;

[0015] Figure 5 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0017] In one embodiment, such as Figure 1 As shown, a method for detecting weak targets based on speed gear search is provided, including the following steps:

[0018] Step 102: Obtain the candidate target point set of the current batch of binarized image sequences.

[0019] The candidate target point set is obtained by selecting pixels that have passed the binarization segmentation threshold in the last frame of the binarized image sequence. For example, selecting... common The binarized image sequence of frames is then taken as follows. Points in the frame image that pass the binarization segmentation threshold are used as target candidate points.

[0020] Step 104: Obtain the set of target search speed levels.

[0021] The target search speed level set includes multiple target search speed levels. This set is calculated based on the maximum speed of the target to be detected and the target search step size, where the target speed is measured in pixels per frame. The maximum speed of the target to be detected is prior knowledge. When the target is moving at a constant velocity in a straight line, since its current speed is unknown in advance, this method sets the target search step size based on the known maximum speed, resulting in multiple target search speed levels. Each speed level corresponds to a range of possible current target speeds.

[0022] Step 106: For each candidate target point in the candidate target point set, search within the corresponding search window of each frame of the binary image excluding the last frame at each target search speed level, and count the cumulative number of image frames searched for each candidate target point at each target search speed level.

[0023] The cumulative image frame count is the total number of binarized image frames within the search window that contain pixels that pass the binarization segmentation threshold. For example, for a candidate target point, a specific target search speed setting is selected to traverse the current batch. Frame binarized image sequence, if If a pixel in the search window of the binarized image exceeds the binarization segmentation threshold, the frame count is 1; otherwise, it is 0. Accumulating these counts yields the search speed level corresponding to that target. Frame count value.

[0024] Step 108: Select the maximum cumulative image frame count from the multiple cumulative image frame counts of each candidate target point, and determine whether the corresponding candidate target point is a target point based on whether the maximum cumulative image frame count meets the preset conditions.

[0025] It can be seen that each candidate target point corresponds to multiple target search speed levels. That is, it is unknown beforehand whether a candidate target point is the target to be detected, nor is the actual speed of the candidate target point and the target to be detected known. Therefore, multiple target search speed levels are set for each candidate target point. Based on each target search speed level, searches are performed in the binarized images of all frames except the last binarized image. The maximum cumulative image frame count indicates that the corresponding candidate target point has the highest probability of moving at a given target search speed level. Therefore, the target search speed level corresponding to the maximum cumulative image frame count is taken as the movement speed of the candidate target point. The preset condition for target determination can be: if the maximum cumulative image frame count is not less than a preset frame count decision threshold, then the corresponding candidate target point is the target to be detected, and the corresponding target search speed level is the movement speed range of the target to be detected. In summary, this method does not require a highly accurate estimation of target motion information throughout the process, thus possessing high flexibility.

[0026] The aforementioned method for detecting weak targets based on speed-level search includes: firstly, acquiring a set of candidate target points and a set of target search speed levels for the current batch of binarized image sequences, wherein the set of target search speed levels is calculated based on the maximum motion speed of the target to be detected and the target search step, with the motion speed of the target to be detected in pixels per frame; for each candidate target point in the candidate target point set, searching within the corresponding search window of each frame of binarized image excluding the last frame at each target search speed level, and calculating the cumulative number of image frames searched for each candidate target point at each target search speed level; selecting the maximum cumulative number of image frames from the multiple cumulative number of image frames for each candidate target point, and determining whether the corresponding candidate target point is a target point based on whether the maximum cumulative number of image frames meets a preset condition. For targets moving at a constant velocity in a straight line, it is not necessary to know the target's accurate true velocity in advance. Only its maximum speed is used as a constraint, and an appropriate target search speed level is set. Each candidate target point is searched and counted within the corresponding search window at each target search speed level. Compared with methods based on multiple frames of images that require a precise target motion model, this method greatly reduces its dependence on prior knowledge and is more applicable.

[0027] In one embodiment, the maximum speed of the target to be detected includes Maximum speed of motion in the direction and Maximum velocity in the direction of motion;

[0028] The steps for calculating the target search speed level set based on the maximum speed of the target to be detected and the target search step include:

[0029] Design the target search step based on the maximum speed of the target to be detected;

[0030] according to The maximum velocity in the direction and the target search step are calculated to obtain multiple Directional speed gear boundary values, according to The maximum velocity in the direction and the target search step are calculated to obtain multiple Directional speed gear boundary values;

[0031] According to adjacent Construct multiple direction speed gear boundary value pairs The directional half-open and half-closed speed range, based on multiple The speed range of half-open and half-closed direction and Multiple direction speed gear boundary values ​​of 0 were constructed to obtain multiple Target search speed settings; among which The upper points of the half-open, half-closed velocity range are adjacent. The absolute value of the directional speed gear boundary value is relatively large. Directional speed gear boundary values;

[0032] According to adjacent Construct multiple direction speed gear boundary value pairs The directional half-open and half-closed speed range, based on multiple The speed range of half-open and half-closed direction and Multiple direction speed gear boundary values ​​of 0 were constructed to obtain multiple Target search speed settings; among which The upper points of the half-open, half-closed velocity range are adjacent. The absolute value of the directional speed gear boundary value is relatively large. Directional speed gear boundary values;

[0033] The target search speed levels are calculated by combining the target search speed levels in each X direction and each target search speed level in each Y direction.

[0034] For example, the target to be detected is in the target is The maximum movement speed in the direction is 2 pixels / frame, and the target search step is 0.2, that is, the step is 0.2 pixels / frame between -2 and 2 pixels / frame. The direction can be selected at 21 speed levels: [-2, -1.8], [-1.8, -1.6], ..., [-0.2, 0], 0, (0, 0.2], ..., (1.6, 1.8], (1.8, 2], while simultaneously satisfying the requirement of approximating the speed from the maximum speed. The same logic applies in the direction of search. As can be seen in this example, excluding the target search speed level 0, each target search speed level corresponds to an open and closed interval, with the interval containing the larger absolute speed value being the closed interval. This design ensures no overlap between speed levels, preventing both repetition and omission during search calculations. Compared to having every speed level as a closed interval, this reduces unnecessary redundant calculations and significantly improves search speed. It's worth noting that the speed level division method in this example is only one of the preferred methods. In actual implementation, the opening and closing methods of intervals and the target search step size can be designed according to requirements to obtain the desired set of target search speed levels.

[0035] In one embodiment, the size of the search window is calculated based on the total number of frames of the current batch of binarized images and the target search step.

[0036] The initial size of the search window is calculated based on the total number of frames in the current batch of binarized images and the target search step size.

[0037]

[0038] in, For the size of the search window, This is the rounding function. This represents the total number of frames in the current batch of binarized images. For target search steps, For MATLAB operators;

[0039] When the initial size is not less than the first preset size, the initial size is used as the size of the corresponding search window; when the initial size is less than the first preset size, the preset size is used as the size of the corresponding search window. For example, when the first preset size is 3, the following adjustments are made based on the calculated search window size:

[0040]

[0041] In one embodiment, for each candidate target point in the candidate target point set, a search is performed within the corresponding search window of each of the other binarized frames (excluding the last binarized frame) at each target search speed level, including:

[0042] Based on the pixel position of each candidate target point in the candidate target point set, calculate the motion boundary position of each candidate target point in each of the other frame binarized images (excluding the last frame binarized image) when searching for targets at each target search speed level.

[0043] When the initial size of the search window is not less than the first preset size, the search window is constructed from the corresponding motion boundary position toward the pixel position of the corresponding candidate target point in the binarized image of the last frame with the initial size to perform target search;

[0044] When the initial size of the search window is smaller than the first preset size and larger than the second preset size, the search window is expanded from the corresponding motion boundary toward the pixel position of the corresponding candidate target point in the binarized image of the last frame to obtain the first expanded search window, and the target search is performed within the first expanded search window; the size of the first expanded search window is the first preset size.

[0045] When the initial size of the search window is not greater than the second preset size, the search window is expanded with the corresponding motion boundary position as the center to obtain a second expanded window, and the target search is performed within the second expanded window; the size of the second expanded search window is the first preset size.

[0046] For example, if a batch of 16 image sequences is selected for weak target detection, then the maximum number of pixels that can be moved in the forward recursion process of a target moving at a maximum speed of 2 pixels / frame is 30 pixels. For a target moving at a speed of (1.8, 2] in the direction of the target, the search window sizes for frames 1 to 15 of the 16-frame image are as follows: 3, 3, 3, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 0, 0. Windows with a search pane size greater than or equal to 3×3 remain unchanged. The search window sizes are determined by the target's position inwards from the motion boundary of that frame (towards the target position in frame 16, the same applies below). The search proceeds in the direction of the corresponding search window, searching for pixels within that window. For panes with a size of 2×2, the search is changed to 3×3, and a 3×3 image window is searched inwards. For search windows smaller than or equal to 1, a 3×3 image window is constructed centered on the corresponding target motion boundary position, with the target position at the center of the 3×3 image window. In other words, the process of forward recursive search window counting and decision-making is handled in two ways:

[0047] Scenario 1: direction( (Similar direction, same below) A certain speed gear is in When the frame is not corrected, the window size is greater than or equal to 2. Direction relative to Frame point Search window coordinates for:

[0048]

[0049] Scenario 2: In a certain speed gear of the direction If the window size was 1 before frame correction and then corrected to 3, then... Direction relative to Frame point Search window coordinates for:

[0050]

[0051] like Figure 2 As shown, a schematic diagram of the search window corresponding to each frame of the binarized image is provided when the target search level is (1.8,2].

[0052] In this example, the first preset size is 3×3, and the second preset size is 1×1. (Combined) There are a total of 21×21=441 speed gears, and the search window for each gear is determined in the same way as above during the recursive process.

[0053] It can be seen that for different candidate target points, as long as the target search speed level is the same, the initial size of the search window is also the same for each batch. In a sequence of binarized images, the search window position moves according to the pixel position of the candidate target points, the selected target search speed level, and the frame number of the binarized image in the current batch. Therefore, given the maximum speed of the target to be detected, and after determining the target search speed level set and the number of frames in each batch of binarized image sequences, the search window can be invoked on demand using offline calculation and loading, ensuring real-time target detection to a certain extent. Furthermore, this method employs sequential execution and batch image processing, ensuring high image processing speed and high data throughput.

[0054] In one embodiment, before acquiring the current batch of binarized image sequences, the following is included:

[0055] Obtain the original image sequence;

[0056] Set the image binarization segmentation threshold based on the false alarm rate of a single frame:

[0057]

[0058] in, This refers to the false alarm rate per frame. For binarization segmentation threshold, This is the rounding function. It is the inverse function of the Q function;

[0059] The original image sequence is binarized according to the image binarization segmentation threshold to obtain a binarized image sequence:

[0060]

[0061] in, This is a binary image sequence. This is the original image sequence.

[0062] In one embodiment, the method further includes:

[0063] After detecting the candidate target points in the current batch of binarized image sequences, the overall binarized image sequence is slid according to a preset number of sliding frames to obtain the candidate target point set for the next batch of binarized image sequences. This target search process is repeated until all binarized images participate in the target detection process. Figure 3 As shown, this provides a schematic diagram of image batch processing when the sliding frame number is 1 frame.

[0064] This invention provides a weak target detection method based on speed gear search. The hardware implementation uses basic circuits such as a decision unit, adder, and counter, which has the advantages of simple hardware implementation. The buffer space of infrared images and various calculators can be time-division multiplexed, which has certain advantages in terms of storage and computing power.

[0065] It should be understood that, although Figure 1 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 1 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0066] In one embodiment, such as Figure 4 As shown, a weak target detection device based on speed gear search is provided, including: a candidate target point set acquisition module 402, a speed gear set acquisition module 404, a target search module 406, and a candidate target point determination module 408, wherein:

[0067] The candidate target point set acquisition module 402 is used to acquire the candidate target point set of the current batch of binarized image sequences; the candidate target point set is obtained by selecting pixels that have passed the binarization segmentation threshold in the last frame of the binarized image sequence.

[0068] The speed gear set acquisition module 404 is used to acquire the target search speed gear set; the target search speed gear set includes multiple target search speed gears; the target search speed gear set is calculated based on the maximum motion speed of the target to be detected and the target search step; the motion speed of the target to be detected is in pixels / frame;

[0069] The target search module 406 is used to search for each candidate target point in the candidate target point set within its respective search window in the binarized images of the frames other than the last frame binarized image, at each target search speed level, and to count the cumulative number of image frames searched for each candidate target point at each target search speed level; the cumulative number of image frames is the total number of binarized images in the search window containing pixels that have passed the binarization segmentation threshold.

[0070] The candidate target point determination module 408 is used to select the maximum cumulative image frame count value from the multiple cumulative image frame count values ​​of each candidate target point, and determine whether the corresponding candidate target point is a target point based on whether the maximum cumulative image frame count value meets the preset conditions.

[0071] For specific limitations regarding the weak target detection device based on speed gear search, please refer to the limitations of the weak target detection method based on speed gear search mentioned above, and will not be repeated here. Each module in the aforementioned weak target detection device based on speed gear search can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0072] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 5 As shown, the computer device includes a processor, memory, network interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used to communicate with external terminals via a network connection. When executed by the processor, the computer program implements a method for detecting weak targets based on speed-level search. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0073] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0074] In one embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the method described in the above embodiments.

[0075] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in the above embodiments.

[0076] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0077] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0078] The above embodiments merely illustrate several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for detecting weak targets based on speed gear search, characterized in that, The method includes: Obtain a candidate target point set for the current batch of binarized image sequences; the candidate target point set is obtained by selecting pixels that have passed the binarization segmentation threshold in the last frame of the binarized image sequence. Obtain a set of target search speed levels; the set of target search speed levels includes multiple target search speed levels; the set of target search speed levels is calculated based on the maximum movement speed of the target to be detected and the target search step; the movement speed of the target to be detected is in pixels per frame; For each candidate target point in the candidate target point set, a search is performed in the corresponding search window of each frame of binarized image excluding the last frame binarized image at each target search speed level. The cumulative number of image frames searched for each candidate target point at each target search speed level is counted. The cumulative number of image frames is the total number of binarized image frames in the search window that contain pixels that have passed the binarization segmentation threshold. The maximum cumulative image frame count is selected from the multiple cumulative image frame counts of each candidate target point, and the corresponding candidate target point is determined to be a target point based on whether the maximum cumulative image frame count meets a preset condition.

2. The method according to claim 1, characterized in that, The maximum speed of the target to be detected includes Maximum speed of motion in the direction and Maximum speed of motion in the direction; The steps for calculating the target search speed level set based on the maximum speed of the target to be detected and the target search step include: Design the target search step based on the maximum speed of the target to be detected; according to The maximum velocity in the direction and the target search step are calculated to obtain multiple Directional speed gear boundary values, according to The maximum velocity in the direction and the target search step are calculated to obtain multiple Directional speed gear boundary values; According to adjacent Construct multiple direction speed gear boundary value pairs The directional half-open and half-closed speed range, based on the aforementioned multiple The speed range of half-open and half-closed direction and Multiple direction speed gear boundary values ​​of 0 were constructed to obtain multiple Target search speed settings; among which The upper points of the half-open, half-closed velocity range are adjacent. The absolute value of the directional speed gear boundary value is relatively large. Directional speed gear boundary values; According to adjacent Construct multiple direction speed gear boundary value pairs The directional half-open and half-closed speed range, based on the aforementioned multiple The speed range of half-open and half-closed direction and Multiple direction speed gear boundary values ​​of 0 were constructed to obtain multiple Target search speed settings; among which The upper points of the half-open, half-closed velocity range are adjacent. The absolute value of the directional speed gear boundary value is relatively large. Directional speed gear boundary values; The target search speed levels are calculated by combining the target search speed levels in each X direction and each target search speed level in each Y direction.

3. The method according to claim 1, characterized in that, The size of the search window is calculated based on the total number of frames of the current batch of binarized images and the target search step. The initial size of the search window is calculated based on the total number of frames in the current batch of binarized images and the target search step size. in, For the size of the search window, This is the rounding function. This represents the total number of frames in the current batch of binarized images. For target search steps, For MATLAB operators; When the initial size is not less than the first preset size, the initial size is used as the size of the corresponding search window; When the initial size is smaller than the first preset size, the preset size is used as the size of the corresponding search window.

4. The method according to claim 3, characterized in that, For each candidate target point in the candidate target point set, a search is performed within the corresponding search window of each of the other binarized frames (excluding the last binarized frame) at each target search speed level, including: Based on the pixel position of each candidate target point in the candidate target point set, calculate the motion boundary position of each candidate target point in each of the other frame binarized images (excluding the last frame binarized image) when performing target search at each of the target search speed levels. When the initial size of the search window is not less than the first preset size, the search window is constructed from the corresponding motion boundary position toward the pixel position of the corresponding candidate target point in the binarized image of the last frame with the initial size to perform target search; When the initial size of the search window is smaller than the first preset size and larger than the second preset size, the search window is expanded from the corresponding motion boundary toward the pixel position of the corresponding candidate target point in the binarized image of the last frame to obtain a first expanded search window, and the target search is performed within the first expanded search window; the size of the first expanded search window is the first preset size; When the initial size of the search window is not greater than the second preset size, the search window is expanded with the corresponding motion boundary position as the center to obtain a second expanded window, and the target search is performed within the second expanded window; the size of the second expanded search window is the first preset size.

5. The method according to claim 1, characterized in that, Before obtaining the current batch of binarized image sequences, the following steps are included: Obtain the original image sequence; Set the image binarization segmentation threshold based on the false alarm rate of a single frame; The original image sequence is binarized according to the image binarization segmentation threshold to obtain a binarized image sequence.

6. The method according to claim 5, characterized in that, An image binarization segmentation threshold is set based on the false alarm rate of a single frame. The original image sequence is then binarized according to the image binarization segmentation threshold to obtain a binarized image sequence, including: The image binarization segmentation threshold is set based on the single-frame false alarm rate as follows: in, This refers to the false alarm rate per frame. For binarization segmentation threshold, This is the rounding function. It is the inverse function of the Q function; The original image sequence is binarized according to the image binarization segmentation threshold to obtain the following binarized image sequence: in, This is a binary image sequence. This is the original image sequence.

7. The method according to claim 1, characterized in that, The method further includes: After the candidate target point set in the current batch of binarized image sequence is detected, the overall binarized image sequence is slid according to the preset sliding frame number to obtain the candidate target point set of the next batch of binarized image sequence. The target search process is repeated until all binarized images participate in the target detection process.

8. A weak target detection device based on speed gear search, characterized in that, The device includes: The candidate target point set acquisition module is used to acquire the candidate target point set of the current batch of binarized image sequences; the candidate target point set is obtained by selecting pixels that have passed the binarization segmentation threshold in the last frame binarized image of the binarized image sequence; A speed gear set acquisition module is used to acquire a target search speed gear set; the target search speed gear set includes multiple target search speed gears; the target search speed gear set is calculated based on the maximum movement speed of the target to be detected and the target search step; the movement speed of the target to be detected is in pixels / frame. The target search module is used to search for each candidate target point in the candidate target point set within the corresponding search window of each frame of binary images (excluding the last frame of binary images) at each target search speed level, and to count the cumulative number of image frames searched for each candidate target point at each target search speed level; the cumulative number of image frames is the total number of binary images within the search window containing pixels that have passed the binarization segmentation threshold. The candidate target point determination module is used to select the maximum cumulative image frame count value from the multiple cumulative image frame count values ​​of each candidate target point, and determine whether the corresponding candidate target point is a target point based on whether the maximum cumulative image frame count value meets a preset condition.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images