Underwater target identification method and device based on adaptive matching filter

Abstract

The invention relates to the technical field of target recognition, and provides an underwater target recognition method and device based on a self-adaptive matched optical filter, and the method comprises the steps: obtaining an original underwater image, and carrying out the data preprocessing of the original underwater image, and obtaining an underwater image; performing multi-scale feature extraction on the underwater image to obtain features of an underwater target to be detected; determining a matched optical filter template based on an adaptive threshold and a dynamic adjustment mechanism according to the features of the underwater target to be detected; and processing the underwater image based on the matched light filter template, and identifying an underwater target. According to the method, the defects that underwater target detection is low in accuracy and is easily influenced by background illumination change in the prior art are overcome, and the functions of self-adapting to the background illumination change and accurately identifying the underwater target are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of target recognition, and in particular to an underwater target recognition method and device based on an adaptive matching filter. Background Art

[0002] Deepwater subsea production systems use subsea completion technology combined with fixed platforms, floating production platforms and other facilities to form different forms of offshore oilfield development. A typical subsea production system consists of subsea equipment and surface control facilities. According to the function, it can be divided into wellhead and Christmas tree system, manifold system and connection system, subsea control and umbilical cable system. Underwater oil and gas production can avoid the influence of harsh sea surface conditions such as wind, waves, currents, icebergs, floating ice and ships, and the oil production equipment is on the seabed where the conditions are relatively stable.

[0003] With the continuous development and widespread application of underwater production systems, while underwater production systems are more convenient, they also increase the demand for intelligent technologies in risk assessment and detection and monitoring. Detection is difficult, and there is an urgent need for intelligent detection algorithms, which puts higher requirements on the safety and efficiency of underwater operations. For example, when deepwater underwater production systems are abnormal, scene features such as oil and gas leakage will change with the seasons, ocean currents and light. Conventional submersibles and divers will ignore a lot of information during detection, which can easily lead to missed diagnoses and high misdiagnosis rates.

[0004] Machine vision technology, as an important means of safety monitoring of underwater production systems, is increasingly used in underwater environments. However, the complexity and particularity of the underwater environment, especially the instability of background lighting, brings many challenges to machine vision technology in target detection. In such an environment, accurately identifying and tracking targets is crucial to ensuring the safety and efficiency of underwater production systems.

[0005] At present, underwater machine vision systems mainly rely on matched filter algorithms for target recognition and tracking, but this algorithm is extremely sensitive to background lighting changes when processing small features (such as edges, line segments, corners, etc.), which can easily lead to target loss, missed detection or false alarms, seriously affecting the performance of underwater production systems. More importantly, traditional machine vision methods have limited adaptability to complex backgrounds and are easily affected by background interference, resulting in false detection or missed detection. Summary of the invention

[0006] The present invention provides an underwater target recognition method and device based on an adaptive matched filter, which are used to solve the defects of low underwater target detection accuracy and susceptibility to background illumination changes in the prior art, and realize the functions of adaptive background illumination changes and accurate recognition of underwater targets.

[0007] The present invention provides an underwater target recognition method based on an adaptive matched filter, comprising the following steps: Acquire an original underwater image, and perform data preprocessing on the original underwater image to obtain an underwater image; Performing multi-scale feature extraction on the underwater image to obtain features of the underwater target to be detected; According to the characteristics of the underwater target to be detected, determining a matching filter template based on an adaptive threshold and a dynamic adjustment mechanism; The underwater image is processed based on the matched filter template to identify underwater targets.

[0008] According to the underwater target recognition method based on adaptive matched filters as claimed in claim 1, the multi-scale feature extraction of the underwater image to obtain the underwater target features to be detected specifically includes: constructing a multi-scale representation of the underwater image; extracting features of underwater images of each scale through a matched filter algorithm; fusing features extracted from different scales to obtain a feature set; performing feature selection on the feature set to remove redundant and irrelevant information; generating a target feature library based on the feature set obtained after feature selection, and using the features in the target feature library as the underwater target features to be detected.

[0009] According to the underwater target recognition method based on adaptive matching filter provided by the present invention, the matching filter template is determined based on the characteristics of the underwater target to be detected, based on an adaptive threshold and a dynamic adjustment mechanism, specifically including: dynamically determining the parameters of the matching filter according to the characteristics of the underwater target to be detected based on the dynamic adjustment mechanism; determining the adaptive threshold of the matching filter according to the background lighting conditions of the underwater image; and determining the matching filter template according to the parameters and the adaptive threshold.

[0010] According to the underwater target recognition method based on adaptive matching filters provided by the present invention, the underwater image is processed based on the matching filter template to identify the underwater target, which specifically includes: sliding the matching filter template on the underwater image; at each sliding window position, determining the degree of feature matching between the matching filter template and the local area of ​​the underwater image in the sliding window; when the feature matching degree exceeds a preset feature matching threshold, determining that there is an underwater target in the local area of ​​the underwater image, and marking the identified underwater target.

[0011] According to the underwater target recognition method based on adaptive matching filters provided by the present invention, after processing the underwater image based on the matching filter template, the method also includes: obtaining the processing result of the underwater image, and determining the feedback result of the matching filter template according to the processing result; optimizing the matching filter template according to the feedback result to determine the optimal parameter configuration.

[0012] According to the underwater target recognition method based on adaptive matched filters provided by the present invention, after recognizing the underwater target, the method further includes: marking the underwater target in the current image with a marking frame; for subsequent images of the current image in the video frame, calculating a characteristic value according to background statistical features in the area to be tested in the subsequent images; comparing the current characteristic value with the adaptive threshold, and when the characteristic value exceeds the adaptive threshold, identifying the area to be tested as the underwater target and marking it with a marking frame.

[0013] An underwater target recognition device based on an adaptive matched filter includes the following modules: An image preprocessing module is used to obtain an original underwater image and perform data preprocessing on the original underwater image to obtain an underwater image; A multi-scale feature extraction module, used to perform multi-scale feature extraction on the underwater image to obtain features of the underwater target to be detected; A matched filter template determination module, used to determine a matched filter template based on an adaptive threshold and a dynamic adjustment mechanism according to the characteristics of the underwater target to be detected; The underwater target recognition module is used to process the underwater image based on the matched filter template to recognize the underwater target.

[0014] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the underwater target recognition method based on the adaptive matching filter as described above is implemented.

[0015] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the underwater target recognition method based on the adaptive matched filter as described in any one of the above is implemented.

[0016] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the method for underwater target recognition based on an adaptive matched filter as described above is implemented.

[0017] The present invention provides an underwater target recognition method and device based on an adaptive matching filter. By performing data preprocessing on the original underwater image, the noise in the image can be effectively removed, the image quality can be improved, and clearer and more accurate image data can be provided for subsequent target recognition. The multi-scale feature extraction method can capture the features of the underwater target at different scales, thereby more comprehensively describing the characteristics of the target and improving the accuracy of recognition. By determining the matching filter template through an adaptive threshold and a dynamic adjustment mechanism, the filter parameters can be automatically adjusted to adapt to different underwater environments and target characteristics, thereby enhancing the flexibility and accuracy of target recognition. By processing the underwater image based on the matching filter template, the target can be effectively identified from a complex underwater environment, thereby improving the efficiency and accuracy of target recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 It is a flow chart of the underwater target recognition method based on the adaptive matched filter provided by the present invention.

[0020] Figure 2 It is a structural schematic diagram of an underwater target recognition device based on an adaptive matched filter provided by the present invention.

[0021] Figure 3 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] The deepwater underwater production system was developed in the 1960s. It uses underwater completion technology combined with fixed platforms, floating production platforms and other facilities to form different forms of offshore oilfield development. A typical underwater production system consists of underwater equipment and surface control facilities. According to the function, it can be divided into wellhead and Christmas tree system, manifold system and connection system, underwater control and umbilical cable system. Underwater oil and gas production can avoid the influence of harsh sea surface conditions such as wind, waves, currents, icebergs, floating ice and ships, and the oil production equipment is on the seabed with relatively stable conditions. While the underwater production system is convenient, it also increases the demand for intelligent technology for risk assessment and detection and monitoring. The detection is difficult and the demand for intelligent detection algorithms is urgent. For example, when the deepwater underwater production system is abnormal, the scene characteristics such as oil and gas leakage will change with the changes of seasons, ocean currents and light. Conventional submersibles and divers will ignore a lot of information during inspection, resulting in missed diagnosis and high misdiagnosis rate.

[0024] In underwater environments, changes in background lighting and water quality may cause the target features to be blurred or hidden. Traditional machine vision-based methods have limited adaptability to complex backgrounds and are easily affected by background interference, resulting in false detection or missed detection.

[0025] In order to solve the above problems and the defects in the background technology, the present invention proposes an improved matched filter algorithm. The algorithm optimizes the design of the filter so that it can better adapt to changes in background lighting when processing small features such as edges, line segments, and corners. Specifically, the present invention introduces an adaptive threshold and a dynamic adjustment mechanism so that the algorithm can adjust its detection parameters according to the real-time environment, thereby improving the accuracy and stability of target detection. Therefore, the present invention, based on an optical, non-contact adaptive matched filter, can well solve the bottleneck problem of missed diagnosis and high misdiagnosis rate caused by the above-mentioned conventional examinations.

[0026] The present invention aims to solve the problems of target loss, missed detection and false alarm caused by characteristic changes in background illumination when a machine vision system uses a matched filter algorithm to detect a target in an underwater environment with unstable background illumination.

[0027] Combine the following Figure 1-Figure 3 Embodiments of the present invention are described.

[0028] Figure 1 FIG. 1 is one of the flow charts of the underwater target recognition method based on the adaptive matched filter provided by the present invention, such as Figure 1 As shown, the method includes the following: S110, acquiring an original underwater image, and performing data preprocessing on the original underwater image to obtain an underwater image.

[0029] According to the underwater target recognition method based on adaptive matched filter provided by the present invention, data preprocessing includes denoising, contrast enhancement and color correction.

[0030] Specifically, the original underwater images of underwater scenes are acquired through underwater cameras, laser scanners and underwater imaging equipment, and the equipment is ensured to be correctly set up to obtain high-quality original underwater images, and the captured original underwater images are stored in appropriate data structures for subsequent processing.

[0031] The stored raw underwater image data is imported into the image processing system to check the integrity and quality of the image data and to remove damaged or low-quality images.

[0032] Perform denoising, contrast enhancement, color correction, etc. on the original underwater image: The original underwater image is denoised by using denoising algorithms, such as median filtering, Gaussian filtering, etc., to reduce noise interference in the image and improve image quality.

[0033] Contrast enhancement: Use histogram equalization, contrast stretching and other technologies to enhance the contrast of the original underwater image and make the target features more obvious.

[0034] Color correction: Since the underwater environment may cause color distortion, color correction is required to restore the true color of the image, such as through white balance adjustment, color space conversion, etc.

[0035] You can also crop and zoom the image appropriately as needed to highlight the target area or adjust the image size.

[0036] The quality of the preprocessed image is evaluated to ensure that the preprocessing step does not introduce too much distortion or deformation. The preprocessed image is used as the underwater image, which will be used for subsequent feature extraction, target recognition and other steps.

[0037] Store the preprocessed underwater images in appropriate data structures for subsequent analysis and processing.

[0038] Since underwater targets are generally in motion and underwater image acquisition equipment is also in motion, data preprocessing of the original underwater images can significantly improve image quality, reduce noise interference, highlight target features, and restore true colors, providing a higher quality data foundation for subsequent feature extraction and target recognition, thereby improving the accuracy and efficiency of underwater target recognition.

[0039] S120, performing multi-scale feature extraction on the underwater image to obtain features of the underwater target to be detected.

[0040] According to the underwater target recognition method based on adaptive matched filters provided by the present invention, multi-scale feature extraction is performed on underwater images to obtain features of underwater targets to be detected, which specifically includes: constructing a multi-scale representation of underwater images; extracting features of underwater images of each scale through a matched filter algorithm; fusing features extracted from different scales to obtain a feature set; performing feature selection on the feature set to remove redundant and irrelevant information; generating a target feature library based on the feature set obtained after feature selection, and using the features in the target feature library as the features of underwater targets to be detected.

[0041] Specifically, an image pyramid is constructed for the preprocessed underwater image. The image pyramid is a set of images arranged in a pyramid shape with gradually decreasing resolutions. This is achieved by continuously smoothing and downsampling the original underwater image.

[0042] At each level of the image pyramid, a matched filter algorithm is applied. Since the resolution of each level is different, the matched filter algorithm can capture target features at different scales.

[0043] At each scale, matched filters are used to detect features such as edges, line segments, corners, etc. The features extracted from different scales are fused through weighted averaging, maximum selection or other fusion strategies to ensure that all important features are captured and used for subsequent object detection.

[0044] In the fused feature set, feature selection is performed to remove redundant and irrelevant information while retaining the most beneficial features for target detection. The optimized feature set is integrated into a target feature library for subsequent target detection and recognition tasks.

[0045] By constructing an image pyramid and applying a matched filter algorithm at different scales, the features of the target at each scale can be fully captured and integrated. This multi-scale feature extraction method enhances the richness and robustness of features, especially in complex and changing underwater environments, and helps improve the accuracy and stability of underwater target recognition.

[0046] S130 . Determine a matching filter template according to the characteristics of the underwater target to be detected, based on an adaptive threshold and a dynamic adjustment mechanism.

[0047] According to the underwater target recognition method based on adaptive matching filter provided by the present invention, a matching filter template is determined according to the characteristics of the underwater target to be detected based on an adaptive threshold and a dynamic adjustment mechanism, specifically including: dynamically determining the parameters of the matching filter according to the characteristics of the underwater target to be detected based on the dynamic adjustment mechanism; determining the adaptive threshold of the matching filter according to the background lighting conditions of the underwater image; and determining the matching filter template according to the parameters and the adaptive threshold.

[0048] Specifically, according to the underwater target features to be detected extracted from the preprocessed image data, the initial parameters of the matching filter are set, including the size, shape, direction of the filter and the threshold used for feature matching.

[0049] Analyze the background lighting conditions of underwater images and dynamically set an initial adaptive threshold accordingly. This threshold will be used in the subsequent feature matching process to ensure effective target recognition under different lighting conditions. The adaptive threshold can be adjusted based on image properties such as brightness and contrast to ensure good detection results under different lighting conditions.

[0050] A dynamic adjustment algorithm is designed to adjust the parameters of the matching filter according to the statistical characteristics of the target features and the expected target appearance.

[0051] Using the initialized parameters and adaptive thresholds, one or more matching filter templates are created. These matching filter templates will be used to search for areas in the underwater image that match the underwater target features to be detected.

[0052] By dynamically determining the parameters of the matching filter according to the characteristics of the underwater target to be detected, and determining the adaptive threshold in combination with the background lighting conditions of the underwater image, and then generating a matching filter template, the matching filter can be more flexibly and accurately adapted to different underwater environments and target characteristics, thereby improving the accuracy and robustness of underwater target recognition.

[0053] S140: Process the underwater image based on the matched filter template to identify the underwater target.

[0054] According to the underwater target recognition method based on adaptive matching filters provided by the present invention, underwater images are processed based on matching filter templates to identify underwater targets, specifically including: sliding the matching filter template on the underwater image; at each sliding window position, determining the degree of feature matching between the matching filter template and the local area of ​​the underwater image in the sliding window; when the feature matching degree exceeds a preset feature matching threshold, determining that an underwater target exists in the local area of ​​the underwater image, and marking the identified underwater target.

[0055] Specifically, the designed matched filter template (referred to as template) is applied to the preprocessed underwater image, and the matched filter template is slid on the image to find the area matching the template.

[0056] At each sliding window position, the matching degree between the template and the local area of ​​the image is determined by calculating the correlation, similarity or other matching metrics between the template and the image area.

[0057] A feature matching threshold is set to determine whether the matching degree of a certain area is high enough to determine whether a target is recognized. When the matching degree of a certain area exceeds the set feature matching threshold, it is considered that an underwater target is recognized in the area. The area is marked, for example, with a rectangular frame, ellipse or other shape to mark the target position.

[0058] If there are multiple targets in the underwater image, a unique identifier needs to be assigned to each identified target. By associating the identified targets between consecutive frames with the unique identifier, their positions and motion states can be tracked. The identified underwater target information is output, including the target's location, size, shape and other attributes. At the same time, this information is recorded for subsequent analysis and processing.

[0059] By matching the filter template to search the underwater image comprehensively, the matching degree between the template and the local area of ​​the image can be accurately evaluated, and the target can be marked when the matching degree exceeds the preset threshold. This method can efficiently identify the target in the underwater image, improve the accuracy and reliability of recognition, and provide strong support for underwater detection, monitoring and other applications.

[0060] According to the underwater target recognition method based on adaptive matching filter provided by the present invention, after processing the underwater image based on the matching filter template, the processing result of the underwater image is obtained, and the feedback result of the matching filter template is determined according to the processing result; the matching filter template is optimized according to the feedback result to determine the optimal parameter configuration.

[0061] Specifically, the recognition results are evaluated, and the accuracy and stability of the recognition results are calculated to evaluate the performance of the template, including the calculation of indicators such as recognition accuracy and recall rate. According to the evaluation results, the matching filter template or recognition algorithm is optimized and adjusted to improve the recognition performance.

[0062] Furthermore, the dynamic adjustment algorithm designed above can optimize the current detection parameters based on the results of the previous detection based on the feedback mechanism. For example, if there are many missed detections or false detections in the previous detection, the algorithm can automatically adjust the threshold or other parameters of the filter to improve the performance. Feedback on the detection results is continuously collected, and the matched filter template is optimized in real time using the dynamic adjustment mechanism.

[0063] The dynamic adjustment mechanism based on feedback results can continuously improve the accuracy and stability of underwater target recognition, enabling the system to gradually reach optimal performance through continuous learning and optimization, thereby better adapting to the complex and changing underwater environment.

[0064] According to the underwater target recognition method based on adaptive matched filters provided by the present invention, after recognizing the underwater target, the method further includes: marking the underwater target in the current image with a marking frame; for subsequent images of the current image in the video frame, calculating a characteristic value according to background statistical features in the area to be tested in the subsequent images; comparing the current characteristic value with the adaptive threshold, and when the characteristic value exceeds the adaptive threshold, identifying the area to be tested as the underwater target and marking it with a marking frame.

[0065] Specifically, after the underwater target is identified, a marking box (such as a rectangular box) is used to mark the target in the image. For subsequent images of the current image in the video frame, the areas to be tested are selected. These areas can be predicted positions based on the target position in the previous frame, or potential target areas determined by an algorithm (such as motion detection). Then, the system calculates feature values ​​based on the background statistical features of these areas to be tested.

[0066] The characteristic value calculated based on the background statistical characteristics can be calculated using the following formula: Eigenvalue=μ + k * σ * (1 + λ * background_clutter)

[0067] Where μ is the background brightness mean, σ is the background brightness standard deviation, k is the basic threshold coefficient, and λ is the adaptive coefficient. Both k and λ are constants and need to be optimized and debugged according to the actual scene. The value of background_clutter can be determined by calculating the mean and standard deviation of the pixel gradient amplitude in the background area.

[0068] The calculated eigenvalue is compared with the adaptive threshold. If the eigenvalue exceeds the adaptive threshold, it means that the area to be tested is likely to contain underwater targets, and it is identified as an underwater target and marked with a marking box.

[0069] Through the above method, when the position of the underwater target in the current image changes to the next detection scene, the adaptive matched filter can be seamlessly connected. At the same time, it has dynamic predictive capabilities in the previous detection scene and can predictively perceive and locate the upcoming underwater target.

[0070] The solution of the present invention can significantly improve the accuracy and stability of target detection in underwater production system safety monitoring when dealing with an underwater environment with unstable background lighting. Specifically, the present invention can effectively reduce the situations of target loss, missed detection and false alarm caused by changes in background lighting, thereby improving the safety and efficiency of underwater production systems.

[0071] When the underwater target is in motion and the underwater image acquisition device is also in motion, the adaptive matched filter proposed in the present invention can dynamically eliminate the relative error generated when both parties move, and adaptively compensate for the distance difference, phase difference and reference change when there is relative motion between the underwater target and the underwater image acquisition device.

[0072] The underwater target recognition device based on the adaptive matched filter provided by the present invention is described below. The underwater target recognition device based on the adaptive matched filter described below and the underwater target recognition method based on the adaptive matched filter described above can be referred to each other.

[0073] like Figure 2 The figure shows a schematic diagram of the structure of an underwater target recognition device based on an adaptive matched filter provided by the present invention, which includes the following modules: The image preprocessing module 210 is used to obtain the original underwater image and perform data preprocessing on the original underwater image to obtain the underwater image; A multi-scale feature extraction module 220 is used to extract multi-scale features from underwater images to obtain features of underwater targets to be detected; A matched filter template determination module 230, configured to determine a matched filter template based on an adaptive threshold and a dynamic adjustment mechanism according to the characteristics of the underwater target to be detected; The underwater target recognition module 240 is used to process the underwater image based on the matched filter template to recognize the underwater target.

[0074] Figure 3 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 3 As shown, the electronic device may include: a processor 310, a communication interface 320, a memory 330 and a communication bus 340, wherein the processor 310, the communication interface 320 and the memory 330 communicate with each other through the communication bus 340. The processor 310 may call the logic instructions in the memory 330 to execute an underwater target recognition method based on an adaptive matched filter, the method comprising: acquiring an original underwater image, performing data preprocessing on the original underwater image to obtain an underwater image; performing multi-scale feature extraction on the underwater image to obtain features of an underwater target to be detected; determining a matched filter template based on an adaptive threshold and a dynamic adjustment mechanism according to the features of the underwater target to be detected; processing the underwater image based on the matched filter template to identify the underwater target.

[0075] In addition, the logic instructions in the above-mentioned memory 330 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0076] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the underwater target recognition method based on adaptive matching filters provided by the above methods. The method includes: acquiring an original underwater image, performing data preprocessing on the original underwater image to obtain an underwater image; performing multi-scale feature extraction on the underwater image to obtain features of an underwater target to be detected; determining a matching filter template based on an adaptive threshold and a dynamic adjustment mechanism according to the features of the underwater target to be detected; and processing the underwater image based on the matching filter template to identify the underwater target.

[0077] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, it is implemented to execute the underwater target recognition method based on adaptive matching filters provided by the above-mentioned methods, the method comprising: acquiring an original underwater image, performing data preprocessing on the original underwater image to obtain an underwater image; performing multi-scale feature extraction on the underwater image to obtain features of an underwater target to be detected; determining a matching filter template based on an adaptive threshold and a dynamic adjustment mechanism according to the features of the underwater target to be detected; and processing the underwater image based on the matching filter template to identify the underwater target.

[0078] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, i.e., they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative effort.

[0079] Through the description of the above implementation modes, those skilled in the art can clearly understand that each implementation mode can be implemented by means of software plus a necessary general hardware platform, or of course by hardware. Based on such an understanding, the above technical solution can essentially or in other words be embodied in the form of a software product that contributes to the prior art. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods of each embodiment or some parts of the embodiment.

[0080] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An underwater target recognition method based on adaptive matched filter, characterized in that: include: Acquire an original underwater image, and perform data preprocessing on the original underwater image to obtain an underwater image; Performing multi-scale feature extraction on the underwater image to obtain features of the underwater target to be detected; According to the characteristics of the underwater target to be detected, determining a matching filter template based on an adaptive threshold and a dynamic adjustment mechanism; The underwater image is processed based on the matched filter template to identify underwater targets.

2. The underwater target recognition method based on adaptive matched filter according to claim 1, characterized in that: The performing multi-scale feature extraction on the underwater image to obtain the underwater target features to be detected specifically includes: constructing a multi-scale representation of the underwater image; The underwater images at each scale are subjected to feature extraction through matched filter algorithm; The features extracted from different scales are fused to obtain a feature set; Performing feature selection on the feature set to remove redundant and irrelevant information; A target feature library is generated according to the feature set obtained after feature selection, and the features in the target feature library are used as underwater target features to be detected.

3. The underwater target recognition method based on adaptive matched filter according to claim 1, characterized in that: The step of determining a matching filter template based on the characteristics of the underwater target to be detected and based on an adaptive threshold and a dynamic adjustment mechanism specifically includes: Based on a dynamic adjustment mechanism, dynamically determining the parameters of the matching filter according to the characteristics of the underwater target to be detected; determining an adaptive threshold of a matching filter according to background lighting conditions of the underwater image; A matched filter template is determined according to the parameters and the adaptive threshold.

4. The underwater target recognition method based on adaptive matched filter according to claim 1, characterized in that: The processing of the underwater image based on the matched filter template to identify the underwater target specifically includes: Sliding the matched filter template on the underwater image; At each sliding window position, determining a degree of feature matching between the matching filter template and a local area of ​​the underwater image in the sliding window; When the feature matching degree exceeds a preset feature matching threshold, it is determined that there is an underwater target in the local area of ​​the underwater image, and the identified underwater target is marked.

5. The underwater target recognition method based on adaptive matched filter according to claim 1, characterized in that: After processing the underwater image based on the matched filter template, the method further includes: Acquiring a processing result of the underwater image, and determining a feedback result of the matching filter template according to the processing result; The matched filter template is optimized according to the feedback result to determine the best parameter configuration.

6. The underwater target recognition method based on adaptive matched filter according to claim 1, characterized in that: After identifying the underwater target, the method further includes: Mark the underwater target in the current image with a marking frame; For subsequent images of the current image in the video frame, characteristic values ​​are calculated according to background statistical features in the area to be tested in the subsequent images; The current feature value is compared with the adaptive threshold, and when the feature value exceeds the adaptive threshold, the area to be detected is identified as the underwater target and marked with a marking frame.

7. An underwater target recognition device based on adaptive matched filter, characterized in that: include: An image preprocessing module is used to obtain an original underwater image and perform data preprocessing on the original underwater image to obtain an underwater image; A multi-scale feature extraction module, used to perform multi-scale feature extraction on the underwater image to obtain features of the underwater target to be detected; A matched filter template determination module, used to determine a matched filter template based on an adaptive threshold and a dynamic adjustment mechanism according to the characteristics of the underwater target to be detected; The underwater target recognition module is used to process the underwater image based on the matched filter template to recognize the underwater target.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the underwater target recognition method based on the adaptive matched filter as described in any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the underwater target recognition method based on the adaptive matched filter as claimed in any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the underwater target recognition method based on the adaptive matched filter as claimed in any one of claims 1 to 6 is implemented.

Images