A method and system for underwater target tracking based on multi-frame forward-looking sonar images

Abstract

This invention relates to the field of underwater target detection and information processing technology, and in particular to an underwater target tracking method and system based on multi-frame forward-looking sonar images. The invention includes: processing a sequence of forward-looking sonar images; identifying suspected targets using a target detection model and calculating the latitude and longitude coordinates of the suspected targets in the geodetic coordinate system; estimating the physical size of the suspected targets based on the detection results; generating measurement objects; for each existing track containing suspected targets, calculating an adaptive correlation gate radius based on navigation system errors, sonar measurement errors, and the average physical size of the targets tracked by the track; establishing a correlation gate; filtering out measurement points in the current frame that fall within the correlation gate to form an effective measurement set; constructing a cost matrix between the track and the measurement points in the effective measurement set; and solving the cost matrix using the Hungarian algorithm to obtain the globally optimal track-measurement correlation relationship. This invention significantly improves tracking continuity and accuracy.

Description

Technical Field

[0001] This invention relates to the field of underwater target detection and information processing technology, and in particular to an underwater target tracking method and system based on multi-frame forward-looking sonar images. Background Technology

[0002] Forward-looking sonar is a key sensor for detecting underwater targets. It is an active sonar that points the acoustic transmitting and receiving arrays directly in front of the carrier and uses fan-shaped / matrix beams to image the water in front of it in real time.

[0003] The core challenge in existing underwater target tracking technologies lies in: 1) The lack of target kinematic information causes the prediction module of the traditional filtering algorithm to fail; 2) The movement of the carrier causes a change in the target's position in the image; 3) Sonar images have a low signal-to-noise ratio and a high false alarm rate; 4) As an extended target, the detection center point of a mine will naturally drift within its physical outline. Traditional correlation gates based on pure sensor error are difficult to effectively contain this drift, which can easily lead to track breakage. Summary of the Invention

[0004] In view of the shortcomings of the prior art described above, the purpose of this invention is to provide an underwater target tracking method and system based on multi-frame forward-looking sonar images, which significantly improves tracking continuity and accuracy by introducing an adaptive adjustment correlation gate for the target's physical size.

[0005] To achieve the above and other related objectives, the present invention provides an underwater target tracking method based on multi-frame forward-looking sonar images, comprising: S1. Measurement Generation and Target Size Estimation: Process forward-looking sonar image sequences, identify suspected targets through a target detection model, calculate the latitude and longitude coordinates of the suspected targets in the geodetic coordinate system, estimate the physical size of the suspected targets based on the detection results, and generate measurement objects. S2. Adaptive correlation gate calculation and gate selection: For each track where the suspected target already exists, an adaptive correlation gate radius is calculated based on the navigation system error, sonar measurement error, and the average physical size of the target tracked by the track. The correlation gate is established, and the measurement points of the current frame that fall into the correlation gate are selected to form an effective measurement set. S3. Data Association: Construct a cost matrix between the track and the measurement points in the effective measurement set, and use the Hungarian algorithm to solve the cost matrix to obtain the globally optimal track-measurement association relationship.

[0006] In one embodiment of the present invention, it further includes: S4. Track Update and Management: Based on the aforementioned association, update the position, image features, average physical size, and occurrence counter of the matching track; update the loss counter of the unmatched track; initialize a new track for the unmatched measurement; and perform track confirmation and deletion management based on the values ​​of the occurrence counter and the loss counter.

[0007] In one embodiment of the present invention, the measurement objects include latitude and longitude, confidence level, image feature vector, and physical size estimate; the establishment of the correlation gate includes establishing the correlation gate with the latest position of the track as the center and the radius as the range.

[0008] In one embodiment of the present invention, the cost matrix integrates normalized geographical distance and image feature similarity.

[0009] In one embodiment of the present invention, the estimated physical size of the suspected target in step S1 is calculated based on the pixel size of the suspected target detection box, the current range of the sonar, and the horizontal beamwidth.

[0010] In one embodiment of the present invention, calculating an adaptive correlation gate radius in step S2 includes: The adaptive correlation gate radius The calculation formula is: , in, For the horizontal positioning error of the navigation system, This is the overall measurement error of the sonar system. A coverage factor between 3 and 5. For the track The historical average physical radius.

[0011] In one embodiment of the present invention, the average physical size in step S4 The exponential moving average method is used for updating, and the formula is as follows: , in, Forgetting factor, This is the estimated physical size carried in the current matching measurement.

[0012] In one embodiment of the present invention, the cost matrix between the track and the measurement points in the effective measurement set in step S3 is calculated as follows:

[0013] in, For the track With measurement Geographical distance between them For the track The radius of the associated gate, For the track The average image feature vector, For measurement Image feature vectors, and The weighting coefficients and .

[0014] In one embodiment of the present invention, step S4, which involves managing the confirmation and deletion of flight tracks based on the values ​​of the occurrence counter and the loss counter, includes: When a flight path appears, the counter Reaching or exceeding the confirmation threshold When this happens, the track status is set to confirmed; When a track loss counter value Reaching or exceeding the deletion threshold When this happens, the track will be deleted from the system.

[0015] The present invention also provides an underwater target tracking system based on multi-frame forward-looking sonar images, comprising: The measurement generation and target size estimation module is used to process forward-looking sonar image sequences, identify suspected targets through a target detection model, calculate the latitude and longitude coordinates of the suspected targets in the geodetic coordinate system, estimate the physical size of the suspected targets based on the detection results, and generate measurement objects. The adaptive correlation gate calculation and gate selection module is used to calculate an adaptive correlation gate radius for each existing track of the suspected target, based on navigation system error, sonar measurement error and the average physical size of the target tracked by the track, establish a correlation gate, and filter out the current frame measurement points that fall into the correlation gate to form an effective measurement set. The data association module is used to construct the cost matrix between the track and the measurement points in the effective measurement set, and to solve the cost matrix using the Hungarian algorithm to obtain the globally optimal track-measurement association relationship.

[0016] As described above, the underwater target tracking method and system based on multi-frame forward-looking sonar images of the present invention have the following beneficial effects: The underwater target tracking method based on multi-frame forward-looking sonar images of the present invention utilizes multi-frame forward-looking sonar images to perform fusion operations based on latitude and longitude and image features, and uses the physical size of the target as one of the core parameters for association gate calculation, so that the tracking model is more in line with the physical reality of the underwater target and solves the problem of association failure caused by the natural drift of the detection point.

[0017] The underwater target tracking method based on multi-frame forward-looking sonar images of the present invention no longer has a fixed correlation gate size, but is adaptively adjusted according to the size of the target being tracked on a specific trajectory, thus achieving fine-grained processing.

[0018] The underwater target tracking method based on multi-frame forward-looking sonar images of the present invention significantly reduces the risk of multiple detections of the same extended target being misidentified as different targets, greatly reduces track breakage, and improves tracking robustness.

[0019] The underwater target tracking method based on multi-frame forward-looking sonar images of the present invention combines a multi-frame confirmation mechanism with feature matching, which can effectively distinguish between real targets and cluttered interference. Attached Figure Description

[0020] Figure 1 A flowchart illustrating an underwater target tracking method based on multi-frame forward-looking sonar images provided in this application embodiment.

[0021] Figure 2 shows the actual tracking results of an underwater target tracking method based on multi-frame forward-looking sonar images provided in the embodiments of this application. (a) is the start frame of sonar target tracking, (b) is the middle frame of sonar target tracking, and (c) is the end frame of sonar target tracking. Detailed Implementation

[0022] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, unless otherwise specified, the following embodiments and features described therein can be combined with each other.

[0023] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the illustrations only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0024] Terms such as "first" or "second" may be used to describe various components, but these components are not limited by the terms described above. The terms described above are used to distinguish one component from another; for example, without departing from the scope of the concept according to this disclosure, a first component may be referred to as a second component, and similarly, a second component may be referred to as a first component.

[0025] Furthermore, "connected / linked" indicates that one component is directly electrically connected to another component or indirectly electrically connected through another component. Unless otherwise explicitly stated in the sentence, the singular form may include the plural form. Additionally, the terms "comprising / including" or "containing / including" as used in this specification indicate the presence or addition of one or more components, steps, operations, and elements. Specific structural or functional descriptions of examples of embodiments of the concepts disclosed in this specification are merely illustrative to describe examples of embodiments of the concepts, and examples of embodiments of the concepts can be implemented in various forms, but these descriptions are not limited to the examples of embodiments described in this specification.

[0026] Based on the concept, various modifications and changes can be applied to examples of embodiments, such that examples of embodiments will be illustrated in the accompanying drawings and described in the specification. However, examples of embodiments based on the concept are not limited to specific embodiments, but include all changes, equivalents, or substitutions included within the spirit and scope of this disclosure.

[0027] It should be understood that when describing an element as "connected" or "linked" to another element, the element may be directly connected or linked to the other element, or it may be connected or linked to the other element via a third element. Conversely, it should be understood that when an element is described as "directly connected to" or "directly linked to" another element, no other element is placed between them. Other expressions describing relationships between components (i.e., "between" and "directly between" or "adjacent to" and "directly adjacent to") need to be interpreted in the same way.

[0028] The terminology used in this specification is for the purpose of describing specific examples of implementations only and is not intended to limit this disclosure. The singular form may include the plural form unless there is an explicit contrary meaning in the context. It should be understood in this specification that the terms "comprising" or "having" indicate the presence of the features, quantities, steps, operations, components, parts, or combinations thereof described in the specification, but do not preclude the possibility of the presence or addition of one or more other features, quantities, steps, operations, components, parts, or combinations thereof.

[0029] Unless otherwise defined, all terms used herein (including technical or scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art. If a term is not clearly defined in a common dictionary in this specification, it shall be interpreted as having the same meaning as in the context of the relevant art, and not as an ideal or overly formal meaning.

[0030] Descriptions of known components and processing techniques may be omitted to avoid unnecessarily obscuring the embodiments of this disclosure.

[0031] Throughout this specification, the same reference numerals refer to the same elements. Therefore, even if a reference numeral is not mentioned or described with reference to one drawing, it may be mentioned or described with reference to another drawing. Furthermore, even if a reference numeral is not shown in one drawing, it may be mentioned or described with reference to another drawing.

[0032] Additionally, the logic level of a signal may be different from or opposite to the logic level described. For example, a signal described as having a logic "high" level may optionally have a logic "low" level, and a signal described as having a logic "low" level may optionally have a logic "high" level.

[0033] The embodiments of this disclosure will now be described in detail with reference to the accompanying drawings. However, those skilled in the art will understand that many technical details have been provided in the embodiments of this disclosure to facilitate a better understanding of the disclosure. However, the technical solutions claimed in this disclosure can be implemented even without these technical details and various variations and modifications based on the following embodiments.

[0034] Please see Figure 1 , Figure 1 A flowchart illustrating an underwater target tracking method based on multi-frame forward-looking sonar images, provided as an embodiment of this application. This invention provides an underwater target tracking method based on multi-frame forward-looking sonar images, comprising: S1. Measurement Generation and Target Size Estimation: Process forward-looking sonar image sequences, identify suspected targets through a target detection model, calculate the latitude and longitude coordinates of the suspected targets in the geodetic coordinate system, estimate the physical size of the suspected targets based on the detection results, and generate measurement objects. S2. Adaptive correlation gate calculation and gate selection: For each track where the suspected target already exists, an adaptive correlation gate radius is calculated based on the navigation system error, sonar measurement error, and the average physical size of the target tracked by the track. The correlation gate is established, and the measurement points of the current frame that fall into the correlation gate are selected to form an effective measurement set. S3. Data Association: Construct a cost matrix between the track and the measurement points in the effective measurement set, and use the Hungarian algorithm to solve the cost matrix to obtain the globally optimal track-measurement association relationship.

[0035] S4. Track Update and Management: Based on the aforementioned association, update the position, image features, average physical size, and occurrence counter of the matching track; update the loss counter of the unmatched track; initialize a new track for the unmatched measurement; and perform track confirmation and deletion management based on the values ​​of the occurrence counter and the loss counter.

[0036] This invention provides an underwater target tracking method based on multi-frame forward-looking sonar images. This method uses multi-frame forward-looking sonar images to perform fusion operations based on latitude and longitude and image features, incorporates the target's physical size as a key parameter into the tracking process, and clarifies the complete link from target measurement, adaptive correlation gate calculation, data correlation to track management, thereby achieving high-reliability, low-false-alarm-rate continuous tracking and confirmation of underwater targets.

[0037] Specifically, the measurement objects include latitude and longitude, confidence level, image feature vector, and physical size estimate; the association gate is established by using the latest track position as the center and a radius as the range. The cost matrix integrates normalized geographical distance and image feature similarity.

[0038] Specifically, the estimated physical size of the suspected target in step S1 is calculated based on the pixel size of the suspected target detection box, the current range of the sonar, and the horizontal beamwidth.

[0039] Specifically, calculating an adaptive association gate radius in step S2 includes: The adaptive correlation gate radius The calculation formula is: , in, For the horizontal positioning error of the navigation system, This is the overall measurement error of the sonar system. A coverage factor between 3 and 5. For the track The historical average physical radius.

[0040] Specifically, the average physical size in step S4 The exponential moving average method is used for updating, and the formula is as follows: , in, Forgetting factor, This is the estimated physical size carried in the current matching measurement.

[0041] Specifically, in step S3, the cost matrix between the track and the measurement points in the effective measurement set is constructed using the following formula:

[0042] in, For the track With measurement Geographical distance between them For the track The radius of the associated gate, For the track The average image feature vector, For measurement Image feature vectors, and The weighting coefficients and .

[0043] Specifically, step S4, based on the values ​​of the occurrence counter and the loss counter, performs track confirmation and deletion management, including: When a flight path appears, the counter Reaching or exceeding the confirmation threshold When this happens, the track status is set to confirmed; When a track loss counter value Reaching or exceeding the deletion threshold When this happens, the track will be deleted from the system.

[0044] In one embodiment of the present invention, an underwater target tracking method based on multi-frame forward-looking sonar images includes: S100: Measurement Generation and Target Size Estimation S101. Image Detection: For the k-th frame of the sonar image, a deep learning target detection model is used to identify suspected targets and output their pixel coordinates and detection box size.

[0045] S102. Geographic coordinate calculation: For each detected target, the absolute latitude and longitude coordinates in the geodetic coordinate system are calculated by fusing the carrier navigation information and sonar parameters.

[0046] S103. Physical Size Estimation: Based on the pixel size of the detection frame, the current range of the sonar, and the beamwidth, estimate the physical size of the target (length, width, or circumcircle). ))) S104. Measurement Packaging: Package the solution results into a single measurement object.

[0047]

[0048] in: It is an image feature vector. This is the estimated target radius for this frame.

[0049] S200: Adaptive Associative Gate Calculation and Gate Selection For each existing track in the system (Its state includes the average target radius) ): Calculate the adaptive correlation gate radius:

[0050] in, For the horizontal positioning error of the navigation system, This is the overall measurement error of the sonar system. A coverage factor between 3 and 5. For the track The historical average physical radius.

[0051] Gate selection: based on the latest position of the flight path With the center of the circle, Construct a circular association gate with radius . Calculate the geodetic distance between each measurement point in the current frame and the center of this circle. .like Then the measurement Falling into the track Within the associated gates, it is included in its effective measurement set. .

[0052] S300: Data Association S301. Construct the cost matrix: for all tracks and their valid measurement sets The measurement constructs the cost matrix Matrix elements Cost of using multi-factor fusion:

[0053] S302. Global Optimal Allocation: The cost matrix... Input the Hungarian algorithm, solve for the optimal allocation scheme, and output the set of matching pairs. Unmatched track set and mismatched measurement set .

[0054] S400: Track Update and Management S401. Update Match Track: For each match pair : Status Update: The track's latest position is updated directly using the measured latitude and longitude.

[0055] Feature Update: Features Using Measurements Average characteristics of updated tracks .

[0056] Size update: using measured radius Update the average radius of the track .

[0057] Counter update: Counter appears Increment by 1, missing counter Reset to 0.

[0058] S402. Update Unmatched Tracks: The status of unmatched tracks remains unchanged, and the loss counter is updated. Add 1.

[0059] S403. Initiating a New Track: For The measurements in the middle are used to initialize a new track, with its initial radius. .

[0060] S404. Track Lifecycle Management: Confirmed: If a temporary flight path Then it will be upgraded to a confirmed track.

[0061] Delete: If a track If so, then delete it.

[0062] Please refer to Figure 2. Figure 2 shows the actual tracking results of an underwater target tracking method based on multi-frame forward-looking sonar images provided in this application embodiment. (a) is the start frame of sonar target tracking, (b) is the middle frame of sonar target tracking, and (c) is the end frame of sonar target tracking. The underwater target tracking method based on multi-frame forward-looking sonar images of this invention is set in the following scenario: an unmanned underwater vehicle (UUV) is surveying a suspicious body of water. The sonar frame rate is 1 Hz, and the UUV searches along a lawnmower-like route at a certain speed, resulting in significant variations in the observation angle of the same area.

[0063] System parameter presets: Navigation error , Sonar error

[0064] Coverage factor , Sensor error term .

[0065] Confirmation threshold

[0066] Deletion threshold

[0067] Cost weight

[0068] Target: There are two real targets at the bottom of the waterway: Target A: Large in size, with a typical physical radius of about 1.5m.

[0069] Target B: Smaller in size, with a typical physical radius of about 0.8m.

[0070] At the same time, random false alarms exist in the environment.

[0071] Execution process of the example: Frame 1: S100: Three measurements were detected.

[0072] / / Suspected Target A / / Suspected Target B / / False Alarm S200-S400: No existing tracks in the system. Initialize three temporary tracks T1, T2, and T3 for M1, M2, and M3 respectively.

[0073]

[0074]

[0075]

[0076] Frame 2: S100: Two measurements were detected. (False alarm M3 has disappeared, and a new false alarm has been generated.)

[0077]

[0078] S200: Calculate the adaptive gate radius for each track and perform gate selection.

[0079] Door T1 radius: 14.9m. Calculate the distances to M4 and M5. M4 falls into the doorway, M5 does not.

[0080] T2 door radius: 14.3m. M4 and M5 did not fall into its doorway.

[0081] T3 door radius: 13.7m. M4 and M5 did not fall into its doorway.

[0082] Valid measurement set: , ,

[0083] S300: Construct the cost matrix (mainly the costs of T1 and M4) The Hungarian algorithm matches T1 with M4.

[0084] S400: Update T1: Position updated to M4 coordinates; Feature and average size updated, for example... ; .

[0085] Update T2, T3: Increased to 1.

[0086] Initiating a new track: M5 did not match any track, so it was initialized as a new track T4.

[0087] Frame 3: S100: Three measurements were detected.

[0088] / / Target B Target A / / New false alarm S200: Calculate the door radius and select the door.

[0089] T1 door radius: 14.95m. M7 fell into the door, while M6 and M8 did not.

[0090] T2 gate radius: 14.3m. M6 falls into its gate.

[0091] T3 door radius: 13.7m. M8 fell into its doorway.

[0092] T3 door radius: 14.6m. M7 fell into its doorway.

[0093] Valid measurement set: , , ,

[0094] S300: Calculate C(T1, M7) and C(T4, M7). Since the historical features of T1 are more similar to those of M7, and the average size of T1 matches the estimated size of M7 better, therefore C(T1, M7) is... <C(T4, M7)。

[0095] The final assignments using the Hungarian algorithm are: T1-M7, T2-M6, T3-M8. T4 is not matched.

[0096] S400: Update T1: Position, Features, and Size Update It was confirmed as the real target.

[0097] Update T2: Position updated to M6 coordinates, feature dimensions updated. ,

[0098] Update T3: Position updated to M8 coordinates, feature dimensions updated. ,

[0099] Update T4: No match.

[0100] Frame 4: S100: Three measurements were detected.

[0101] Target A / / Target B S200: After gate selection: After the selection, M9 fell into door T1; M10 fell into door T2; neither M9 nor M10 fell into doors T3 or T4.

[0102] S300: Hungarian algorithm allocation: T1-M9, T2-M10.

[0103] S400: Update T1, T2: Update the status of T2 It was confirmed as the real target.

[0104] Update T2, position updated to M6 coordinates, feature dimensions updated. ,

[0105] Update T3 and T4. Increased to 1 and 2 respectively Frame 5 and onwards: Targets A and B were continuously detected and successfully associated, and their tracks remained confirmed.

[0106] T3 (false alarm) and T4 (false alarm) were not detected in subsequent frames.

[0107] By frame 8, T3's T4 All of them reached the deletion threshold and were deleted by the system.

[0108] Similar in principle to the underwater target tracking method based on multi-frame forward-looking sonar images of the present invention, the present invention also provides an underwater target tracking system based on multi-frame forward-looking sonar images, comprising: The measurement generation and target size estimation module is used to process forward-looking sonar image sequences, identify suspected targets through a target detection model, calculate the latitude and longitude coordinates of the suspected targets in the geodetic coordinate system, estimate the physical size of the suspected targets based on the detection results, and generate measurement objects. The adaptive correlation gate calculation and gate selection module is used to calculate an adaptive correlation gate radius for each existing track of the suspected target, based on navigation system error, sonar measurement error and the average physical size of the target tracked by the track, establish a correlation gate, and filter out the current frame measurement points that fall into the correlation gate to form an effective measurement set. The data association module is used to construct the cost matrix between the track and the measurement points in the effective measurement set, and to solve the cost matrix using the Hungarian algorithm to obtain the globally optimal track-measurement association relationship.

[0109] The track update and management module is used to update the position, image features, average physical size and occurrence counter of matching tracks according to the association relationship, update the loss counter of unmatched tracks, initialize new tracks for unmatched measurements, and perform track confirmation and deletion management based on the values ​​of the occurrence counter and the loss counter.

[0110] In summary, the underwater target tracking method based on multi-frame forward-looking sonar images of the present invention utilizes multi-frame forward-looking sonar images to perform fusion operations based on latitude and longitude and image features, and uses the physical size of the target as one of the core parameters for association gate calculation, making the tracking model more closely match the physical reality of the underwater target and solving the association failure problem caused by the natural drift of the detection point.

[0111] This invention also proposes an electronic device comprising a processor and a memory. The memory stores program instructions, and the processor executes these instructions to implement the aforementioned underwater target tracking method based on multi-frame forward-looking sonar images. The processor can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The memory may include Random Access Memory (RAM) and may also include Non-Volatile Memory, such as at least one disk storage device. The memory can also be an internal memory of the Random Access Memory (RAM) type. The processor and memory can be integrated into one or more independent circuits or hardware, such as an Application Specific Integrated Circuit (ASIC). It should be noted that when the computer program in the aforementioned memory is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, an electronic device, or a network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention.

[0112] This invention also proposes a computer-readable storage medium storing computer instructions for instructing a computer to execute the aforementioned underwater target tracking method based on multi-frame forward-looking sonar images. The computer-readable storage medium can be an electronic medium, magnetic medium, optical medium, electromagnetic medium, infrared medium, or semiconductor system or propagation medium. The computer-readable storage medium may also include semiconductor or solid-state memory, magnetic tape, removable computer disk, random access memory (RAM), read-only memory (ROM), hard disk, and optical disc. Optical discs may include optical disc-read-only memory (CD-ROM), optical disc-read / write (CD-RW), and DVD.

[0113] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. An underwater target tracking method based on multi-frame forward-looking sonar images, characterized in that, include: S1. Measurement Generation and Target Size Estimation: Process forward-looking sonar image sequences, identify suspected targets through a target detection model, calculate the latitude and longitude coordinates of the suspected targets in the geodetic coordinate system, estimate the physical size of the suspected targets based on the detection results, and generate measurement objects. S2. Adaptive correlation gate calculation and gate selection: For each track where the suspected target already exists, an adaptive correlation gate radius is calculated based on the navigation system error, sonar measurement error, and the average physical size of the target tracked by the track. The correlation gate is established, and the measurement points of the current frame that fall into the correlation gate are selected to form an effective measurement set. S3. Data Association: Construct a cost matrix between the track and the measurement points in the effective measurement set, and use the Hungarian algorithm to solve the cost matrix to obtain the globally optimal track-measurement association relationship.

2. The underwater target tracking method based on multi-frame forward-looking sonar images according to claim 1, characterized in that, Also includes: S4. Track Update and Management: Based on the aforementioned association, update the position, image features, average physical size, and occurrence counter of the matching track; update the loss counter of the unmatched track; initialize a new track for the unmatched measurement; and perform track confirmation and deletion management based on the values ​​of the occurrence counter and the loss counter.

3. The underwater target tracking method based on multi-frame forward-looking sonar images according to claim 2, characterized in that: The measurement objects include latitude and longitude, confidence level, image feature vector, and physical size estimate; the establishment of the correlation gate includes establishing the correlation gate with the latest position of the track as the center and the radius as the range.

4. The underwater target tracking method based on multi-frame forward-looking sonar images according to claim 2, characterized in that: The cost matrix integrates normalized geographic distance and image feature similarity.

5. The underwater target tracking method based on multi-frame forward-looking sonar images according to claim 2, characterized in that, The estimated physical size of the suspected target in step S1 is calculated based on the pixel size of the suspected target detection box, the current range of the sonar, and the horizontal beamwidth.

6. The underwater target tracking method based on multi-frame forward-looking sonar images according to claim 2, characterized in that, Calculating an adaptive associative gate radius in step S2 includes: The adaptive correlation gate radius The calculation formula is: ， in, The horizontal positioning error of the navigation system, This is the overall measurement error of the sonar system. A coverage factor between 3 and 5. For the track The historical average physical radius.

7. The underwater target tracking method based on multi-frame forward-looking sonar images according to claim 6, characterized in that, Average physical dimensions in step S4 The exponential moving average method is used for updating, and the formula is as follows: ， in, Forgetting factor, This is the estimated physical size carried in the current matching measurement.

8. The underwater target tracking method based on multi-frame forward-looking sonar images according to claim 7, characterized in that, In step S3, the cost matrix between the track and the measurement points in the effective measurement set is constructed using the following formula: in, For the track With measurement Geographical distance between them For the track The radius of the associated gate, For the track The average image feature vector, For measurement Image feature vectors, and The weighting coefficients and .

9. The underwater target tracking method based on multi-frame forward-looking sonar images according to claim 8, characterized in that, Step S4, based on the values ​​of the occurrence counter and the loss counter, performs track confirmation and deletion management, including: When a flight path appears, the counter Reaching or exceeding the confirmation threshold When this happens, the track status is set to confirmed; When a track loss counter value Reaching or exceeding the deletion threshold When that happens, the track will be deleted from the system.

10. An underwater target tracking system based on multi-frame forward-looking sonar images, characterized in that, include: The measurement generation and target size estimation module is used to process forward-looking sonar image sequences, identify suspected targets through a target detection model, calculate the latitude and longitude coordinates of the suspected targets in the geodetic coordinate system, estimate the physical size of the suspected targets based on the detection results, and generate measurement objects. The adaptive correlation gate calculation and gate selection module is used to calculate an adaptive correlation gate radius for each existing track of the suspected target, based on navigation system error, sonar measurement error and the average physical size of the target tracked by the track, establish a correlation gate, and filter out the current frame measurement points that fall into the correlation gate to form an effective measurement set. The data association module is used to construct the cost matrix between the track and the measurement points in the effective measurement set, and to solve the cost matrix using the Hungarian algorithm to obtain the globally optimal track-measurement association relationship. The track update and management module is used to update the position, image features, average physical size and occurrence counter of matching tracks according to the association relationship, update the loss counter of unmatched tracks, initialize new tracks for unmatched measurements, and perform track confirmation and deletion management based on the values ​​of the occurrence counter and the loss counter.

Images