High-frequency radar time-frequency domain ship target detection method and system based on improved YOLOv5s

By improving the YOLOv5s model, high-frequency radar ship target detection is transformed into an image processing task. The network structure is optimized by using dynamic serpentine convolution and triple attention mechanism, which solves the problem of poor ship target detection performance in high-frequency radar systems and realizes adaptive detection of ship targets.

CN118964932AActive Publication Date: 2024-11-15WUHAN UNIV
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
CN202410999489.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-24
Publication Date
2024-11-15
Estimated Expiration
2044-07-24

AI Technical Summary

Technical Problem

Traditional constant false alarm rate (CFAR) detection algorithms are difficult to effectively detect ship targets in high-frequency radar systems due to fluctuations in the ship's cross-section and velocity. Furthermore, existing time-frequency domain detection methods have poor adaptive performance under high noise conditions.

Method used

An improved YOLOv5s model is adopted to transform high-frequency radar ship target energy detection into a target detection task in image processing. The network structure is optimized through dynamic serpentine convolution and triple attention mechanism, and the time-frequency ridge line is extracted by greedy algorithm to achieve adaptive detection.

Benefits of technology

It improves the effectiveness and feature extraction capabilities of high-frequency radar for ship target detection, simplifies the operation process, and enables adaptive detection of strong, weak, and non-stationary targets.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118964932A_ABST
    Figure CN118964932A_ABST
Patent Text Reader

Abstract

The invention discloses a high-frequency radar time-frequency domain ship target detection method and system based on improved YOLOv5s, and the method comprises the steps: carrying out the calculation according to the obtained high-frequency radar receiving data, obtaining an initial radar detection matrix and a preprocessing time-frequency spectrum, and obtaining the distance element information of a target according to the preprocessing time-frequency spectrum; inputting the preprocessed time-frequency spectrum into a trained improved YOLOv5s model for target detection to obtain a bounding box of a target; adopting a greedy algorithm to perform time-frequency ridge extraction on the bounding box of the target in combination with the corresponding preprocessing time-frequency spectrum, and obtaining the center Doppler frequency of the target according to the extracted time-frequency ridge; and obtaining a final target radar detection matrix according to the distance element and the center Doppler frequency of the target in combination with the initial radar detection matrix. According to the method, the usability of the method can be ensured, and meanwhile, the detection performance of the high-frequency radar on the ship target is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of ship target detection, and in particular to a high-frequency radar time-frequency domain ship target detection method and system based on improved YOLOv5s. Background Art

[0002] High-frequency radar systems can realize continuous monitoring of large areas of sea around the clock and in all weather conditions, and are an important technical means for dynamic monitoring of the marine environment. In the process of using such systems for ship target detection, due to the fluctuations in the ship's scattering cross section and the fluctuations in the speed of travel, the echo of the ship target often shows non-stationary characteristics of Doppler and energy. In this case, it is difficult to obtain ideal detection results using traditional constant false alarm detection algorithms for ship target detection.

[0003] Since the time-frequency analysis method can capture the time-varying characteristics of non-stationary signals, many scholars are committed to studying the methods of ship target detection in the time-frequency domain to improve the target detection performance of high-frequency radar systems. The core idea of ​​the high-frequency radar time-frequency domain detection method is to convert the traditional range-Doppler spectrum midpoint target detection task into a ridge extraction task in the time-frequency spectrum, which is generally designed by combining edge tracking algorithms with energy judgment. However, this type of method requires a large number of parameters to be set in combination with artificial experience to achieve control of the detection process. In the case of variable noise floors in the high-frequency band, the adaptive detection performance of the method is poor. Summary of the invention

[0004] In order to overcome the shortcomings of the above-mentioned prior art, the present invention provides a high-frequency radar time-frequency domain ship target detection method and system based on improved YOLOv5s. By improving the YOLOv5s model, the high-frequency radar ship target energy detection task is converted into a target detection task in image processing, achieving the technical effects of good detection effect, strong feature extraction capability and convenient operation.

[0005] According to one aspect of the present invention, a high-frequency radar time-frequency domain ship target detection method based on improved YOLOv5s is provided, comprising: According to the acquired high-frequency radar reception data, an initial radar detection matrix and a preprocessed time-frequency spectrum are calculated, and the distance meta-information of the target is obtained according to the preprocessed time-frequency spectrum; Input the preprocessed time-frequency spectrum into the trained improved YOLOv5s model to perform target detection to obtain a bounding box of the target; A greedy algorithm is used to extract the time-frequency ridge of the target's bounding box in combination with its corresponding preprocessed time-frequency spectrum, and the center Doppler frequency of the target is obtained according to the extracted time-frequency ridge; According to the target's range element and center Doppler frequency, combined with the initial radar detection matrix, the final target radar detection matrix is ​​obtained; The training of the improved YOLOv5s model includes: Construct a training set of preprocessed time-frequency spectra; The bottleneck structure of YOLOv5s is optimized by using the dynamic snake convolution module and triple attention mechanism module to build an improved YOLOv5s model. The improved YOLOv5s model is trained using the training set of the preprocessed time-frequency spectrum to obtain a trained improved YOLOv5s model.

[0006] As a further technical solution, after obtaining the high-frequency radar receiving data, it also includes: According to the acquired high-frequency radar receiving data, a high-frequency radar distance-slow time spectrum is obtained; According to the high-frequency radar range-slow time spectrum, the range-Doppler spectrum and the time-frequency spectrum of each range element are calculated respectively; According to the dimension of the range-Doppler spectrum, an initial radar detection matrix with the same dimension is generated.

[0007] As a further technical solution, a training set of preprocessed time-frequency spectrum is constructed, including: According to the range-Doppler spectrum, a first-order spectral region Doppler range and a ground clutter Doppler range are calculated respectively; Setting the first-order spectrum area and the ground clutter Doppler area in the time-frequency spectrum to 0; The time-frequency spectrum after the zero-setting operation is normalized to obtain a preprocessed time-frequency spectrum.

[0008] As a further technical solution, the method further includes: dividing the preprocessed time-frequency spectrum into a training set and a test set of the preprocessed time-frequency spectrum according to a preset ratio.

[0009] As a further technical solution, the method further includes: Determine the initial radar detection matrix based on the target's received data field information; According to the distance element information and the center Doppler frequency of the target, the corresponding elements in the initial radar detection matrix are set to 1; The detection results in the initial radar detection matrix are combined in an aggregation manner to obtain a final radar detection matrix.

[0010] As a further technical solution, the method further includes: after a preset time or a preset number of reception sessions, retraining the constructed improved YOLOv5s model using a new preprocessed time-frequency spectrum training set to obtain new pre-trained weights.

[0011] According to one aspect of the present invention, a high-frequency radar time-frequency domain ship target detection system based on improved YOLOv5s is provided, comprising: The first main module is used to calculate the initial radar detection matrix and the preprocessed time-frequency spectrum according to the acquired high-frequency radar reception data, and obtain the distance meta-information of the target according to the preprocessed time-frequency spectrum; The second main module is used to input the preprocessed time-frequency spectrum into the trained improved YOLOv5s model for target detection to obtain a bounding box of the target; The third main module is used to extract the time-frequency ridge of the target's bounding box in combination with its corresponding preprocessed time-frequency spectrum using a greedy algorithm, and obtain the target's central Doppler frequency according to the extracted time-frequency ridge; The fourth main module is used to obtain the final target radar detection matrix based on the target's range element and center Doppler frequency combined with the initial radar detection matrix; The training of the improved YOLOv5s model includes: Construct a training set of preprocessed time-frequency spectra; The bottleneck structure of YOLOv5s is optimized by using the dynamic snake convolution module and triple attention mechanism module to build an improved YOLOv5s model. The improved YOLOv5s model is trained using the training set of the preprocessed time spectrum to obtain pretraining weights, thereby completing the training of the improved YOLOv5s model.

[0012] As a further technical solution, it also includes: a fifth main module, which is used to re-train the constructed improved YOLOv5s model using a new pre-processed time-frequency spectrum training set after a preset time or a preset number of reception sessions to obtain new pre-trained weights.

[0013] According to one aspect of the present specification, there is provided an electronic device, comprising: at least one processor, at least one memory and a communication interface; wherein the processor, memory and communication interface communicate with each other; the memory stores program instructions executed by the processor, and the processor calls the program instructions to execute the described method.

[0014] According to one aspect of the present specification, a non-transitory computer-readable storage medium is provided, wherein the non-transitory computer-readable storage medium stores computer instructions, wherein the computer instructions enable the computer to execute the method described above.

[0015] Aiming at the non-stationary characteristics of ship target echoes in high-frequency radar, the present invention utilizes the advantage of time-frequency domain in capturing the analysis of time-varying signals, and transforms the high-frequency radar ship target energy detection task into a target detection task in image processing. Compared with the prior art, the present invention has the following advantages: The detection effect is good. By marking the time-frequency ridge samples of ship targets with cross-Doppler elements and time-varying amplitude in the training set, adaptive detection of strong targets, weak targets, and non-stationary targets can be achieved. Strong feature extraction capability. By introducing dynamic snake convolution and triple attention mechanism modules, the improved YOLOv5s network has significantly improved its ability to extract time-frequency ridges. The operation is easy. Since the deep learning network is an end-to-end target detection model, it only needs to set a reasonable confidence threshold to achieve control of the target detection task, which is beneficial to improving the performance of high-frequency radar ship target detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, a brief introduction is given below to the drawings used 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 any creative work.

[0017] Figure 1 A flowchart of a high-frequency radar time-frequency domain ship target detection method based on improved YOLOv5s is provided in an embodiment of the present invention.

[0018] Figure 2 A network architecture diagram of the improved YOLOv5s model provided in an embodiment of the present invention.

[0019] Figure 3 A schematic diagram of the CBL module structure in the improved YOLOv5s model provided in an embodiment of the present invention.

[0020] Figure 4 A schematic diagram of the DSCBL module structure in the improved YOLOv5s model provided in an embodiment of the present invention.

[0021] Figure 5 A schematic diagram of the structure of the C3 module in the improved YOLOv5s model provided in an embodiment of the present invention.

[0022] Figure 6 A schematic diagram of the structure of the C3-DSCT module in the improved YOLOv5s model provided in an embodiment of the present invention.

[0023] Figure 7 A schematic diagram of the detection results of the input time-frequency spectrum by the improved YOLOv5 model provided in an embodiment of the present invention.

[0024] Figure 8 A schematic diagram of the result of using a greedy algorithm to extract time-frequency ridges from a bounding box of a predicted target in combination with its corresponding preprocessed time-frequency spectrum provided in an embodiment of the present invention.

[0025] Fig. 9 A schematic diagram of a target parameter conversion process provided by an embodiment of the present invention.

[0026] Fig.10 A schematic diagram of a radar detection matrix without a merging operation provided in an embodiment of the present invention.

[0027] Fig.11 A schematic diagram of a radar detection matrix performing a merging operation provided in an embodiment of the present invention.

[0028] Fig.12 A schematic diagram of the structure of a high-frequency radar time-frequency domain ship target detection system based on improved YOLOv5s provided in an embodiment of the present invention.

[0029] Fig.13 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0030] It should be noted that: In recent years, data-driven deep learning models have shown great potential in tasks such as target detection. Many studies have shown that using a training set with sufficient samples can make deep learning models have superior detection performance. The continuous working ability of high-frequency radar enables the system to obtain rich samples for model training. To this end, the inventor uses the high-frequency radar time-frequency spectrum as the data source, based on the deep learning target detection network YOLOv5s, combined with the greedy algorithm and graph morphological processing technology, to design a detection method suitable for high-frequency radar ship targets, aiming to improve the detection performance of high-frequency radar ship targets.

[0031] The terms "including" and "having" and any variations thereof in the specification and claims of the present invention and the above-mentioned drawings are intended to cover non-exclusive inclusions, for example, a process, method, system, product or apparatus comprising a series of steps or units is not necessarily limited to the steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products or apparatuses.

[0032] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices. The flowcharts shown in the accompanying drawings are merely exemplary and do not necessarily include all the contents and operations / steps, nor do they necessarily have to be executed in the order described. For example, some operations / steps may be further decomposed, while some operations / steps may be combined or partially combined, so the actual execution order may change according to the actual situation.

[0033] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments 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 in 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. In addition, the technical features in the various embodiments or single embodiments provided by the present invention are arbitrarily combined with each other to form a new technical solution. This combination is not restricted by the sequence of steps and / or the structural composition mode, but must be based on the ability of ordinary technicians in this field to achieve. When the combination of technical solutions is contradictory or cannot be achieved, it should be considered that this combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0034] The embodiment of the present invention provides a high-frequency radar time-frequency domain ship target detection method based on improved YOLOv5s, such as Figure 1 As shown, including: Step 1, according to the acquired high-frequency radar reception data, calculate and obtain the initial radar detection matrix and the preprocessing time-frequency spectrum, and obtain the distance meta-information of the target according to the preprocessing time-frequency spectrum; Step 2, inputting the preprocessed time-frequency spectrum into the trained improved YOLOv5s model for target detection to obtain a bounding box of the target; Step 3, using a greedy algorithm to extract the time-frequency ridge of the target's bounding box in combination with its corresponding preprocessed time-frequency spectrum, and obtaining the target's central Doppler frequency according to the extracted time-frequency ridge; Step 4: According to the target's range element and center Doppler frequency, combined with the initial radar detection matrix, the final target radar detection matrix is ​​obtained.

[0035] The training of the improved YOLOv5s model includes: Construct a training set of preprocessed time-frequency spectra; The bottleneck structure of YOLOv5s is optimized by using the dynamic snake convolution module and triple attention mechanism module to build an improved YOLOv5s model. The improved YOLOv5s model is trained using the training set of the preprocessed time-frequency spectrum to obtain a trained improved YOLOv5s model.

[0036] Furthermore, if Figures 2 to 6 As shown in the figure, the dynamic snake convolution module and the triple attention mechanism Triplet Attention module are used to optimize the bottleneck structure of YOLOv5s and build an improved YOLOv5s network model. Then, the improved YOLOv5s model is trained using the constructed time-frequency spectrum training set to obtain the optimal pre-training weights.

[0037] In a high-frequency radar time-frequency domain ship target detection method based on improved YOLOv5s provided in an embodiment of the present invention, Figure 6 As shown in the figure, the dynamic snake convolution is designed for topological tubular structures. By limiting the connectivity of the focus area, the dynamic transformation of the convolution kernel is controlled, and it shows excellent detection performance for continuous tubular structure targets. Given that the target echo of high-frequency radar often appears as a slender and continuous curve in the time-frequency domain, the dynamic snake convolution is used to replace the original two-dimensional convolution module to improve the network's target detection performance in the application scenario of the present invention.

[0038] In a high-frequency radar time-frequency domain ship target detection method based on improved YOLOv5s provided by an embodiment of the present invention, the feature generated by the network of the present invention is Figure 3 The dimensions correspond to the time, Doppler frequency and semantic information dimensions of the target echo respectively. The triple attention mechanism module is a low-computation attention mechanism module. The introduction of this module can help the network learn features. Figure 3 In the scenario of high-frequency radar time-frequency domain target detection, this module helps improve the network's ability to analyze the relationship between the time and Doppler frequency, time and semantics, and semantics and Doppler frequency of the target echo, thereby optimizing the network's detection performance of the target time-frequency ridge.

[0039] As a preferred embodiment, a high-frequency radar time-frequency domain ship target detection method based on improved YOLOv5s provided in an embodiment of the present invention also includes: collecting multiple fields of high-frequency radar receiving data to obtain the high-frequency radar distance-slow time spectrum and receiving data field information.

[0040] Furthermore, an embodiment of the present invention provides a high-frequency radar time-frequency domain ship target detection method based on improved YOLOv5s, such as Fig. 9As shown, it also includes: first, according to the high-frequency radar range-slow time spectrum, the fast Fourier transform method is used to obtain the range-Doppler spectrum; secondly, according to the high-frequency radar range-slow time spectrum, the multiple synchronous squeezing algorithm is used to obtain the time-frequency spectrum of each range element; finally, according to the dimension of the range-Doppler spectrum, a radar detection matrix of the same dimension is generated, and all elements of the matrix are set to 0.

[0041] Preferably, the distance-slow time spectrum is defined as , whose dimension size is Dimension, where N is the maximum distance element to be observed, M is the number of sweep cycles, To receive the data field sequence number, in the embodiment .

[0042] Preferably, the range-Doppler spectrum is defined as , whose dimension size is Dimension, where N is the maximum distance element to be observed, M is the number of sampling points in the Doppler frequency dimension, To receive the data field sequence number, in the embodiment .

[0043] Preferably, the time-frequency spectrum is defined as , whose dimension size is Dimension, where It is the number of sampling points in the time dimension of the time-frequency spectrum, is the number of sampling points of the Doppler frequency dimension of the time-frequency spectrum, To receive the data field sequence number, is the distance element corresponding to the time spectrum, in the embodiment .

[0044] Preferably, the radar detection matrix is ​​defined as , whose dimension size is Dimension, where N is the maximum distance element to be observed, M is the number of sweep cycles, To receive the data field sequence number, in the embodiment .

[0045] As a preferred embodiment, a high-frequency radar time-frequency domain ship target detection method based on improved YOLOv5s provided in an embodiment of the present invention also includes: first, according to the range-Doppler spectrum obtained in the above steps, the Doppler range of the first-order spectrum area is obtained based on the ocean echo Bragg scattering theorem, and the ground clutter Doppler range is obtained based on the stationary characteristics of the ground clutter; secondly, the first-order spectrum area and the ground clutter Doppler area in the obtained time-frequency spectrum are set to 0; and again, the time-frequency spectrum after the setting operation is normalized to obtain a preprocessed time-frequency spectrum.

[0046] Preferably, the pre-processed time spectrum is defined as , whose dimension size is Dimension, where is the number of sampling points in the time dimension of the time-frequency spectrum, is the number of sampling points of the Doppler frequency dimension of the time-frequency spectrum, is the received data field sequence number, n is the distance element corresponding to the time spectrum, in the embodiment .

[0047] Furthermore, during model training, the preprocessed time-frequency spectrum is divided into a training set of the preprocessed time-frequency spectrum and a set to be detected according to a preset ratio, the training set is used to train the improved YOLOv5s model, and the targets in the set to be detected are detected based on the trained improved YOLOv5s model. It should be noted that the preset ratio here can be adaptively adjusted according to actual conditions. After obtaining the training set of the preprocessed time-frequency spectrum, the time-frequency ridge samples of ship targets with cross-Doppler elements and time-varying amplitude can be labeled in the training set to achieve adaptive detection of strong targets, weak targets, and non-stationary targets. Figure 7 As shown, the line is not a straight line and sways left and right, indicating that the target time-frequency ridge crosses the Doppler element, and the brightness of the line is different, indicating that the amplitude of the target time-frequency ridge varies with time.

[0048] It should be noted that after the trained improved YOLOv5s model has been used for a period of time, a new preprocessed time-frequency spectrum can be obtained according to new high-frequency radar reception data, and the model can be retrained using the new preprocessed time-frequency spectrum to obtain a new optimal training weight.

[0049] As a preferred embodiment, a high-frequency radar time-frequency domain ship target detection method based on improved YOLOv5s provided in an embodiment of the present invention further includes: first, according to the pre-processed time-frequency spectrum to be detected, recording the distance meta information and the received data field information of the target; secondly, using the optimal pre-training weight obtained by the previous training, setting the confidence threshold to perform target detection on the pre-processed time-frequency spectrum, and obtaining the bounding box of the predicted target, such as Figure 7 As shown; again, the greedy algorithm is used to extract the time-frequency ridge of the predicted target bounding box combined with its corresponding preprocessed time-frequency spectrum, and then the central Doppler frequency of the target is calculated, as shown in Figure 8 shown.

[0050] Preferably, the confidence threshold is defined as .

[0051] As shown in the figure, the figure shows the detection result of the input time-frequency spectrum by the improved YOLOv5 model, and the red rectangular box is the bounding box of the predicted target, which is generated by YOLOv5 network reasoning; the figure shows the result of using the greedy algorithm to extract the time-frequency ridge of the bounding box of the predicted target in combination with its corresponding pre-processed time-frequency spectrum, and the yellow line is the target time-frequency ridge extracted by the greedy algorithm. The embodiment of the present invention uses the improved YOLOv5s model to predict the bounding box of the target, and then combines the greedy algorithm to perform high-frequency radar time-frequency domain ship target detection, wherein the single time-frequency ridge marked in yellow in the figure swings within a certain Doppler range, and the center Doppler frequency of the target can be obtained by calculating the Doppler mean of this time-frequency ridge.

[0052] As a preferred embodiment, a high-frequency radar time-frequency domain ship target detection method based on improved YOLOv5s provided in an embodiment of the present invention also includes: first, finding the corresponding radar detection matrix according to the obtained received data field information; then, according to the obtained target distance meta-information and the target's central Doppler frequency, setting the corresponding element in the radar detection matrix to 1; finally, using an aggregation method, merging the detection results in the radar detection matrix to obtain a final radar detection matrix.

[0053] Preferably, the final radar detection matrix is ​​defined as , whose dimension size is dimensional, where N is the maximum distance element to be observed, M is the number of sampling points in the Doppler frequency dimension, and k is the sequence number of the received data field. .

[0054] It should be noted that, because in high-frequency radar, targets may be diffuse, the same target in an adjacent area in the radar detection matrix may have multiple detection results, such as Fig.10 Specifically, the distribution of each 8-connected neighborhood in the radar detection matrix can be obtained by using graphic image technology, and then the center of each 8-connected neighborhood is calculated as the detection result of the target, as shown in Fig.11 shown.

[0055] The implementation basis of each embodiment of the present invention is realized by programmed processing through a device with a processor function. Therefore, in engineering practice, the technical solutions and functions of each embodiment of the present invention are encapsulated into various modules. Based on this reality, on the basis of the above embodiments, an embodiment of the present invention provides a high-frequency radar time-frequency domain ship target detection system based on improved YOLOv5s, and the device is used to execute the high-frequency radar time-frequency domain ship target detection method based on improved YOLOv5s in the above method embodiment.

[0056] See also Fig.12The system includes: a first main module, which is used to calculate the initial radar detection matrix and the preprocessed time-frequency spectrum according to the acquired high-frequency radar reception data, and obtain the distance meta-information of the target according to the preprocessed time-frequency spectrum; a second main module, which is used to input the preprocessed time-frequency spectrum into the trained improved YOLOv5s model for target detection and obtain the bounding box of the target; a third main module, which is used to extract the time-frequency ridge of the bounding box of the target in combination with its corresponding preprocessed time-frequency spectrum by using a greedy algorithm, and obtain the central Doppler frequency of the target according to the extracted time-frequency ridge; a fourth main module, which is used to extract the time-frequency ridge of the bounding box of the target according to the extracted time-frequency ridge; and a fourth main module, which is used to extract the time-frequency ridge of the bounding box of the target according to the extracted time-frequency ridge. The module is used to obtain the final target radar detection matrix according to the target range element and the center Doppler frequency in combination with the initial radar detection matrix; wherein the training of the improved YOLOv5s model includes: constructing a training set of preprocessed time-frequency spectrum; optimizing the bottleneck structure of YOLOv5s by using the dynamic snake convolution module and the triple attention mechanism module to construct the improved YOLOv5s model; training the improved YOLOv5s model by using the training set of preprocessed time-frequency spectrum to obtain the pre-training weights, and completing the training of the improved YOLOv5s model.

[0057] The high-frequency radar time-frequency domain ship target detection system based on improved YOLOv5s provided by the embodiment of the present invention aims at the problem that the ship target echo often presents non-stationary characteristics such as Doppler migration and energy fluctuation in the high-frequency radar system. Fig.12 Several modules in the algorithm are used to convert the high-frequency radar ship target energy detection task into the target detection task in image processing based on the analysis advantage of time-frequency domain that can capture time-varying signals. The adaptive detection of strong targets, weak targets and non-stationary targets is realized. The ability of the improved YOLOv5s network in extracting time-frequency ridges is significantly improved, thereby improving the high-frequency radar ship target detection performance.

[0058] It should be noted that the system embodiment provided by the present invention is used to implement the method in the above method embodiment as well as the method in other method embodiments provided by the present invention. The only difference is that the corresponding functional modules are set. The principle is basically the same as the principle of the above system embodiment provided by the present invention. As long as the technical personnel in this field refer to the specific technical solutions in other method embodiments on the basis of the above system embodiment, obtain the corresponding technical means and the technical solutions composed of these technical means by combining technical features, and on the premise of ensuring the practicality of the technical solutions, they can improve the equipment in the above system embodiment to obtain the corresponding system embodiment, which is used to implement the methods in other method embodiments. For example: Based on the content of the above system embodiment, as a preferred embodiment, the high-frequency radar time-frequency domain ship target detection system based on improved YOLOv5s provided in the embodiment of the present invention further includes: A first calculation submodule is used to obtain a high-frequency radar range-slow time spectrum according to the obtained high-frequency radar reception data; The second calculation submodule is used to calculate the range-Doppler spectrum and the time-frequency spectrum of each range element according to the high-frequency radar range-slow time spectrum; The third calculation submodule is used to generate an initial radar detection matrix with the same dimension as the range-Doppler spectrum according to the dimension of the range-Doppler spectrum.

[0059] Based on the content of the above system embodiment, as a preferred embodiment, the high-frequency radar time-frequency domain ship target detection system based on improved YOLOv5s provided in the embodiment of the present invention further includes: A fourth calculation submodule is used to calculate the first-order spectrum region Doppler range and the ground clutter Doppler range according to the range-Doppler spectrum; A fifth calculation submodule, used for setting the first-order spectrum area and the ground clutter Doppler area in the time-frequency spectrum to zero; The sixth calculation submodule is used to perform normalization processing on the time-frequency spectrum after the zero-setting operation to obtain a preprocessed time-frequency spectrum.

[0060] Based on the content of the above system embodiment, as a preferred embodiment, the high-frequency radar time-frequency domain ship target detection system based on improved YOLOv5s provided in the embodiment of the present invention also includes: dividing the preprocessed time-frequency spectrum into a training set and a test set of the preprocessed time-frequency spectrum according to a preset ratio.

[0061] Based on the content of the above system embodiment, as a preferred embodiment, the high-frequency radar time-frequency domain ship target detection system based on improved YOLOv5s provided in the embodiment of the present invention further includes: A seventh calculation submodule is used to determine an initial radar detection matrix according to the received data field information of the target; an eighth calculation submodule, configured to set corresponding elements in the initial radar detection matrix to 1 according to the distance element information and the center Doppler frequency of the target; The ninth calculation submodule is used to merge the detection results in the initial radar detection matrix in an aggregation manner to obtain a final radar detection matrix.

[0062] Based on the content of the above system embodiment, as a preferred embodiment, the high-frequency radar time-frequency domain ship target detection system based on improved YOLOv5s provided in the embodiment of the present invention also includes: after a preset time or a preset number of receiving sessions, the constructed improved YOLOv5s model is re-trained using a new pre-processed time-frequency spectrum training set to obtain new pre-training weights.

[0063] The method of the embodiment of the present invention is implemented by relying on electronic devices, so it is necessary to introduce the relevant electronic devices. Based on this purpose, the embodiment of the present invention provides an electronic device, such as Fig.13 As shown, the electronic device includes: at least one processor, a communication interface, at least one memory, and a communication bus, wherein at least one processor, the communication interface, and at least one memory communicate with each other through the communication bus. At least one processor calls the logic instructions in at least one memory to execute all or part of the steps of the method provided by the aforementioned various method embodiments.

[0064] In addition, when the logic instructions in the above-mentioned at least one memory are implemented in the form of a software functional unit and sold or used as an independent product, they are stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or partly in other words, the part that contributes to the prior art or part of the technical solution is embodied in the form of a software product, and the computer software product is stored in a storage medium, including several instructions for enabling a computer device (a personal computer, a server, or a network device) to perform all or part of the steps of the method described in each method embodiment of the present invention. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a disk or an optical disk, and various media for storing program codes.

[0065] The system embodiments described above are merely illustrative, wherein the units described as separate components are or are not physically separated, and the components shown as units are or are not physical units, located in one place, or distributed on multiple network units. Some or all of the modules are selected according to actual conditions to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement without creative effort.

[0066] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0067] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0068] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0069] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0070] In summary, the present invention obtains the high-frequency radar range-slow time spectrum according to the high-frequency radar receiving data; calculates the range-Doppler spectrum and the time-frequency spectrum of each distance element according to the range-slow time spectrum, and calculates the first-order spectrum and the Doppler range of the ground clutter according to the range-Doppler spectrum; obtains the pre-processed time-frequency spectrum according to the first-order spectrum and the ground clutter range, which is used to prepare the time-frequency target detection data set; uses the time-frequency target detection data set to train the improved YOLOv5s to obtain the pre-training weight; uses the pre-training weight to detect the ridge of the ship target in the time-frequency spectrum; applies the greedy algorithm to obtain the target center Doppler; and finally, based on the image processing technology, maps the time-frequency domain target detection result to the radar detection matrix corresponding to the range-Doppler spectrum. The present invention can improve the detection performance of the high-frequency radar for ship targets while ensuring the ease of use of the method.

[0071] 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 replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the technical solutions of the embodiments of the present invention.

Claims

1. A high-frequency radar time-frequency domain ship target detection method based on improved YOLOv5s is characterized by: include: According to the acquired high-frequency radar reception data, an initial radar detection matrix and a preprocessed time-frequency spectrum are calculated, and the distance meta-information of the target is obtained according to the preprocessed time-frequency spectrum; Input the preprocessed time-frequency spectrum into the trained improved YOLOv5s model to perform target detection to obtain a bounding box of the target; A greedy algorithm is used to extract the time-frequency ridge of the target's bounding box in combination with its corresponding preprocessed time-frequency spectrum, and the center Doppler frequency of the target is obtained according to the extracted time-frequency ridge; According to the target's range element and center Doppler frequency, combined with the initial radar detection matrix, the final target radar detection matrix is ​​obtained; The training of the improved YOLOv5s model includes: Construct a training set of preprocessed time-frequency spectra; The bottleneck structure of YOLOv5s is optimized by using the dynamic snake convolution module and triple attention mechanism module to build an improved YOLOv5s model. The improved YOLOv5s model is trained using the training set of the preprocessed time-frequency spectrum to obtain a trained improved YOLOv5s model.

2. According to the high-frequency radar time-frequency domain ship target detection method based on improved YOLOv5s according to claim 1, it is characterized in that: After obtaining the high-frequency radar receiving data, it also includes: According to the acquired high-frequency radar receiving data, a high-frequency radar distance-slow time spectrum is obtained; According to the high-frequency radar range-slow time spectrum, the range-Doppler spectrum and the time-frequency spectrum of each range element are calculated respectively; According to the dimension of the range-Doppler spectrum, an initial radar detection matrix with the same dimension is generated.

3. According to the high-frequency radar time-frequency domain ship target detection method based on improved YOLOv5s according to claim 2, it is characterized in that: Construct a training set of preprocessed time-frequency spectrograms, including: According to the range-Doppler spectrum, a first-order spectral region Doppler range and a ground clutter Doppler range are calculated respectively; Setting the first-order spectrum area and the ground clutter Doppler area in the time-frequency spectrum to 0; The time-frequency spectrum after the zero-setting operation is normalized to obtain a preprocessed time-frequency spectrum.

4. The high-frequency radar time-frequency domain ship target detection method based on improved YOLOv5s according to claim 3 is characterized in that: The method further includes: dividing the preprocessed time-frequency spectrum into a training set and a test set of the preprocessed time-frequency spectrum according to a preset ratio.

5. The high-frequency radar time-frequency domain ship target detection method based on improved YOLOv5s according to claim 1 is characterized in that: The method further comprises: Determine the initial radar detection matrix based on the target's received data field information; According to the distance element information and the center Doppler frequency of the target, the corresponding elements in the initial radar detection matrix are set to 1; The detection results in the initial radar detection matrix are combined in an aggregation manner to obtain a final radar detection matrix.

6. The high-frequency radar time-frequency domain ship target detection method based on improved YOLOv5s according to claim 1 is characterized in that: The method further includes: after a preset time or a preset number of reception sessions, retraining the constructed improved YOLOv5s model using a new preprocessed time-frequency spectrum training set to obtain new pre-training weights.

7. High-frequency radar time-frequency domain ship target detection system based on improved YOLOv5s, characterized in that: include: The first main module is used to calculate the initial radar detection matrix and the preprocessed time-frequency spectrum according to the acquired high-frequency radar reception data, and obtain the distance meta-information of the target according to the preprocessed time-frequency spectrum; The second main module is used to input the preprocessed time-frequency spectrum into the trained improved YOLOv5s model for target detection to obtain a bounding box of the target; The third main module is used to extract the time-frequency ridge of the target's bounding box in combination with its corresponding preprocessed time-frequency spectrum using a greedy algorithm, and obtain the target's central Doppler frequency according to the extracted time-frequency ridge; The fourth main module is used to obtain the final target radar detection matrix according to the target's range element and center Doppler frequency combined with the initial radar detection matrix; The training of the improved YOLOv5s model includes: Construct a training set of preprocessed time-frequency spectra; The bottleneck structure of YOLOv5s is optimized by using the dynamic snake convolution module and triple attention mechanism module to build an improved YOLOv5s model. The improved YOLOv5s model is trained using the training set of the preprocessed time spectrum to obtain pretraining weights, thereby completing the training of the improved YOLOv5s model.

8. The high-frequency radar time-frequency domain ship target detection system based on improved YOLOv5s according to claim 7 is characterized in that: Also includes: The fifth main module is used to retrain the constructed improved YOLOv5s model using a new pre-processed time-frequency spectrum training set after a preset time or a preset number of reception sessions to obtain new pre-trained weights.

9. An electronic device, characterized in that: include: At least one processor, at least one memory and a communication interface; wherein the processor, memory and communication interface communicate with each other; The memory stores program instructions executed by the processor, and the processor calls the program instructions to execute the method according to any one of claims 1 to 6.

10. A non-transitory computer readable storage medium, characterized in that: The non-transitory computer-readable storage medium stores computer instructions, which cause the computer to execute the method of any one of claims 1 to 6.

Citation Information

Patent Citations

  • Radar interpulse phase coded signal side lobe suppression method

    CN114415119A

  • High-frequency MIMO radar auxiliary calibration source and calibration method

    CN115308703A

  • Non-contact gesture control method and device for transmission system

    CN117636469A

  • Intelligent driving road crack detection method based on dynamic snakelike convolution

    CN117876334A

  • Method for detecting surface defects of electrical insulation paper

    CN118071724A