Multi-mode aviation sonar buoy signal processing and target identification method

Abstract

The invention relates to the technical field of signal target identification, in particular to a multi-mode aviation sonar buoy signal processing and target identification method, which comprises the following steps of: performing signal acquisition in a preset sea area range by using an aviation sonar buoy; the collected multi-mode signals are preprocessed; performing feature extraction on the preprocessed multi-mode signal to generate a unified feature vector; the fused feature vectors are input into a pre-trained target recognition model, and the model adopts a hybrid neural network structure, comprises a convolutional layer and a recurrent neural network layer and is used for processing time sequence signal features; and processing a multi-mode signal input in real time based on the trained target identification model to realize identification and classification of a sea area target, and optimizing an identification result by combining a support vector machine and a classifier of a neural network. According to the method, the continuous capture of the signal sequential relationship is ensured, so that the system can still keep efficient identification and tracking capabilities in a dynamic environment.

Description

Technical Field

[0001] The present invention relates to the technical field of signal target recognition, and in particular to a multi-mode aviation sonar buoy signal processing and target recognition method. Background Art

[0002] With the continuous development of ocean detection and monitoring technology, aviation sonar buoys, as an important detection tool, have been widely used in marine environment monitoring, resource exploration and military purposes, especially in the field of submarine detection. Traditional aviation sonar buoy signal processing methods usually rely on single-mode signal acquisition and processing methods, and common ones are processing technologies based on acoustic signals. However, when facing a complex marine environment, single-mode signal processing methods are often unable to effectively cope with a variety of environmental noise and interference factors, resulting in a decrease in the accuracy and robustness of target recognition. In addition, the signal processing algorithms in the existing technology mostly rely on fixed parameters and lack the ability to adapt to real-time environmental changes, which further limits their application in complex sea conditions.

[0003] In recent years, the rise of multimodal signal processing technology has provided a new approach to solving this problem. By combining different types of signals, such as acoustic signals, vibration signals, and environmental noise signals, and using multimodal data fusion technology, it is possible to mine the complementary information of target features to a greater extent and enhance the accuracy and stability of target recognition. However, the multimodal fusion technology currently on the market still faces the problems of high data processing complexity and poor model adaptability, and cannot achieve ideal recognition effects in complex environments. Summary of the invention

[0004] The invention provides a multi-mode aviation sonar buoy signal processing and target recognition method.

[0005] The multi-mode aeronautical sonar buoy signal processing and target recognition method comprises the following steps:

[0006] S1, multi-mode signal acquisition: Use aviation sonar buoys to collect signals within a predetermined sea area. The collected signals include acoustic signals, vibration signals, and environmental noise signals. Sensor modules of different modes are used to synchronously collect multiple types of signal data to ensure the temporal and spatial consistency of the data.

[0007] S2, signal preprocessing: preprocess the collected multi-mode signals, including denoising, signal enhancement and frequency filtering operations, use adaptive noise suppression algorithm to effectively suppress environmental noise, and extract characteristic signals of different frequency bands through multi-scale wavelet transform method;

[0008] S3, signal feature extraction and fusion: feature extraction of preprocessed multi-modal signals, extraction of high-dimensional features using deep learning-based convolutional neural networks, and dimensionality reduction of redundant features through principal component analysis. Subsequently, multi-modal data fusion technology is used to fuse features of different signal modes to generate a unified feature vector.

[0009] S4, target recognition model training: The fused feature vector is input into the pre-trained target recognition model. The model adopts a hybrid neural network structure, including convolutional layers and recurrent neural network layers, to process time series signal features. Through training on a large-scale target sample data set, the model can achieve accurate recognition and classification of potential targets.

[0010] S5, target recognition and classification: Based on the trained target recognition model, the real-time input multi-modal signals are processed to realize the recognition and classification of sea targets. The support vector machine and neural network classifiers are combined to further optimize the recognition results and improve the accuracy of classification.

[0011] Optionally, the convolutional neural network is used to extract spatial features of multimodal signals, and its network structure includes multiple convolutional layers and pooling layers. The convolutional layer is used to capture local features of the signal, and the pooling layer is used to downsample the features, thereby reducing data dimension and computational complexity. The pooling layer is implemented after each convolutional layer to reduce redundant information while retaining the main features, and improve the efficiency of feature extraction through maximum pooling or average pooling to optimize the speed and accuracy of signal processing.

[0012] Optionally, the hybrid neural network structure combines a convolutional neural network and a recurrent neural network, wherein the convolutional layer is responsible for extracting the spatial features of the signal, and the recurrent neural network is responsible for extracting the timing features of the signal. Through this combined network structure, the spatiotemporal dependencies in multimodal signals can be better processed. The recurrent neural network processes the timing information of the signal, and its hidden layer dynamically updates the state according to historical inputs, so that the system can maintain effective tracking and identification of the target when facing continuous input signals, especially in a dynamically changing environment, ensuring the continuity and accuracy of recognition.

[0013] Optionally, the target recognition and classification process is achieved by combining the use of a support vector machine and a neural network. The support vector machine is used to process high-dimensional features output from the hybrid neural network, and utilizes its sensitivity to boundary data to further improve the classification accuracy, thereby ensuring multi-target recognition and classification capabilities in complex environments.

[0014] Optionally, the support vector machine adopts a nonlinear kernel function to adapt to the nonlinear relationship between different types of signals, thereby improving the classification effect of the system when processing multimodal data. The use of this nonlinear kernel function can enhance the recognition accuracy of complex targets and cope with noise interference in the marine environment.

[0015] Optionally, the network structure of the convolutional neural network includes multiple convolutional layers and pooling layers, wherein the convolution operation is defined as: in, is the convolution output of the lth layer, is the input of the previous layer, is the convolution kernel weight of the lth layer, b (l) is the bias term. The pooling layer downsamples the convolution output to reduce the feature dimension and computational complexity. The pooling operation is defined as: in, Represents the pooling output of layer l, both maximum pooling and average pooling can be applied.

[0016] Optionally, the hybrid neural network structure adopts a combination of a convolutional layer and a recurrent neural network layer, wherein the recurrent neural network layer is used to capture the temporal characteristics of the signal, and the hidden layer state update formula of the recurrent neural network layer is: t =σ(W h h t-1 +W x x t +b h ) where h t is the hidden state at time t, h t-1 is the hidden state of the previous moment, x t is the input at the current moment, W h and W x are weight matrices, b h is the bias term, and σ is the activation function.

[0017] Beneficial effects of the present invention:

[0018] The present invention integrates the features of different types of signals by fusing multi-mode signals, including acoustic signals, vibration signals and environmental noise signals, and utilizes multi-modal data fusion technology, thereby effectively enhancing the system's detection and recognition capabilities for marine targets. Traditional single-mode signal processing methods are easily interfered by environmental noise under complex sea conditions, and the recognition accuracy is not high. The present invention combines an adaptive noise suppression algorithm with a multi-scale wavelet transform, which can effectively suppress environmental noise interference.

[0019] The hybrid neural network structure proposed in the present invention combines a convolutional neural network (CNN) with a recurrent neural network (RNN), which can not only extract the spatial features of the signal, but also capture the timing information of the signal. This structure can fully model the complex spatiotemporal dependencies in multimodal signals, and improve the performance of the system when processing continuous signals, especially in the task of multi-target tracking. The convolutional neural network is used to extract local spatial features, while the recurrent neural network ensures the continuous capture of the timing relationship of the signal, so that the system can still maintain efficient recognition and tracking capabilities in a dynamic environment.

[0020] The present invention further improves the accuracy of target recognition and classification by combining the support vector machine (SVM) and the neural network classifier structure. The support vector machine optimizes the high-dimensional features output by the hybrid neural network, which is particularly suitable for processing complex boundary samples. In addition, the system has real-time feedback and adaptive adjustment functions, and can dynamically adjust the working mode of the buoy, including sensor acquisition frequency, signal processing parameters, etc., according to the results of target recognition and tracking. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0022] Figure 1 The figure is a schematic diagram of a method flow of an embodiment of the present invention. DETAILED DESCRIPTION

[0023] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0024] It should be noted that the references to "one embodiment", "an embodiment", "an exemplary embodiment", "some embodiments" and the like in the specification indicate that the embodiments described may include specific features, structures or characteristics, but not every embodiment may include the specific features, structures or characteristics. In addition, when a specific feature, structure or characteristic is described in conjunction with an embodiment, it should be within the knowledge of a person skilled in the art to implement such feature, structure or characteristic in conjunction with other embodiments (whether or not explicitly described).

[0025] In general, a term can be understood, at least in part, from its use in context. For example, depending, at least in part, on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending, at least in part, on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0026] like Figure 1 As shown, the multi-mode aeronautical sonar buoy signal processing and target recognition method comprises the following steps:

[0027] S1, multi-mode signal acquisition: Use aviation sonar buoys to collect signals within a predetermined sea area. The collected signals include acoustic signals, vibration signals, and environmental noise signals. Sensor modules of different modes are used to synchronously collect multiple types of signal data to ensure the temporal and spatial consistency of the data.

[0028] S2, signal preprocessing: preprocess the collected multi-mode signals, including denoising, signal enhancement and frequency filtering operations, use adaptive noise suppression algorithm to effectively suppress environmental noise, and extract characteristic signals of different frequency bands through multi-scale wavelet transform method;

[0029] S3, signal feature extraction and fusion: feature extraction of preprocessed multi-modal signals, extraction of high-dimensional features using deep learning-based convolutional neural networks, and dimensionality reduction of redundant features through principal component analysis. Subsequently, multi-modal data fusion technology is used to fuse features of different signal modes to generate a unified feature vector.

[0030] S4, target recognition model training: The fused feature vector is input into the pre-trained target recognition model. The model adopts a hybrid neural network structure, including convolutional layers and recurrent neural network layers, to process time series signal features. Through training on a large-scale target sample data set, the model can achieve accurate recognition and classification of potential targets.

[0031] S5, target recognition and classification: Based on the trained target recognition model, the real-time input multi-modal signals are processed to realize the recognition and classification of sea targets. The support vector machine and neural network classifiers are combined to further optimize the recognition results and improve the accuracy of classification.

[0032] Convolutional neural networks are used to extract spatial features of multimodal signals. Their network structure includes multiple convolutional layers and pooling layers. The convolutional layer is used to capture the local features of the signal, while the pooling layer is used to downsample the features, thereby reducing the data dimension and computational complexity. The pooling layer is implemented after each convolutional layer to reduce redundant information while retaining the main features, and improve the efficiency of feature extraction through maximum pooling or average pooling to optimize the speed and accuracy of signal processing.

[0033] The hybrid neural network structure combines the convolutional neural network and the recurrent neural network, in which the convolutional layer is responsible for extracting the spatial features of the signal, and the recurrent neural network is responsible for extracting the temporal features of the signal. Through this combined network structure, the spatiotemporal dependencies in multimodal signals can be better processed. The recurrent neural network processes the timing information of the signal, and its hidden layer dynamically updates the state according to the historical input, so that the system can maintain effective tracking and identification of the target when facing continuous input signals, especially in a dynamically changing environment, ensuring the continuity and accuracy of identification.

[0034] The target recognition and classification process is achieved by combining the use of support vector machines and neural networks. The support vector machine is used to process the high-dimensional features output from the hybrid neural network, and its sensitivity to boundary data is used to further improve the classification accuracy and ensure the ability to recognize and classify multiple targets in complex environments.

[0035] The support vector machine uses a nonlinear kernel function to adapt to the nonlinear relationship between different types of signals, thereby improving the classification effect of the system when processing multimodal data. The use of this nonlinear kernel function can enhance the recognition accuracy of complex targets and cope with noise interference in the marine environment.

[0036] The network structure of the convolutional neural network includes multiple convolutional layers and pooling layers, where the convolution operation is defined as: in, is the convolution output of the lth layer, is the input of the previous layer, is the convolution kernel weight of the lth layer, b (l) is the bias term. The pooling layer downsamples the convolution output to reduce the feature dimension and computational complexity. The pooling operation is defined as: in, Represents the pooling output of layer l, both maximum pooling and average pooling can be applied.

[0037] The hybrid neural network structure adopts a combination of convolutional layer and recurrent neural network layer, where the recurrent neural network layer is used to capture the temporal characteristics of the signal. The hidden layer state update formula of the recurrent neural network layer is: t =σ(W h h t-1 +W x xt +b h ) where h t is the hidden state at time t, h t-1 is the hidden state of the previous moment, x t is the input at the current moment, W h and W x are weight matrices, b h is the bias term, and σ is the activation function.

[0038] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0039] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A multi-mode aerosonar buoy signal processing and target recognition method, characterized in that: The following steps are involved: S1, multi-mode signal acquisition: Use aviation sonar buoys to collect signals within a predetermined sea area. The collected signals include acoustic signals, vibration signals, and environmental noise signals. Sensor modules of different modes are used to synchronously collect multiple types of signal data to ensure the temporal and spatial consistency of the data. S2, signal preprocessing: preprocess the collected multi-mode signals, including denoising, signal enhancement and frequency filtering operations, use adaptive noise suppression algorithm to effectively suppress environmental noise, and extract characteristic signals of different frequency bands through multi-scale wavelet transform method; S3, signal feature extraction and fusion: feature extraction of preprocessed multi-modal signals, extraction of high-dimensional features using deep learning-based convolutional neural networks, and dimensionality reduction of redundant features through principal component analysis. Subsequently, multi-modal data fusion technology is used to fuse features of different signal modes to generate a unified feature vector. S4, target recognition model training: The fused feature vector is input into the pre-trained target recognition model. The model adopts a hybrid neural network structure, including convolutional layers and recurrent neural network layers, to process time series signal features. Through training on a large-scale target sample data set, the model can achieve accurate recognition and classification of potential targets. S5, target recognition and classification: Based on the trained target recognition model, the real-time input multi-modal signals are processed to realize the recognition and classification of sea targets. The support vector machine and neural network classifiers are combined to further optimize the recognition results and improve the accuracy of classification.

2. The multi-mode aeronautical sonar buoy signal processing and target recognition method according to claim 1, characterized in that: The convolutional neural network is used to extract the spatial features of multi-modal signals. Its network structure includes multiple convolutional layers and pooling layers. The convolutional layer is used to capture the local features of the signal, while the pooling layer is used to downsample the features, thereby reducing the data dimension and computational complexity. The pooling layer is implemented after each convolutional layer to reduce redundant information while retaining the main features, and improve the efficiency of feature extraction through maximum pooling or average pooling to optimize the speed and accuracy of signal processing.

3. The multi-mode aeronautical sonar buoy signal processing and target recognition method according to claim 1, characterized in that: The hybrid neural network structure combines a convolutional neural network and a recurrent neural network, wherein the convolutional layer is responsible for extracting the spatial features of the signal, and the recurrent neural network is responsible for extracting the temporal features of the signal. Through this combined network structure, the spatiotemporal dependencies in multimodal signals can be better processed. The recurrent neural network processes the timing information of the signal, and its hidden layer dynamically updates the state according to historical inputs, so that the system can maintain effective tracking and identification of the target when facing continuous input signals, especially in a dynamically changing environment, ensuring the continuity and accuracy of identification.

4. The multi-mode aerosonar buoy signal processing and target recognition method according to claim 1, characterized in that: The target recognition and classification process is achieved by combining the use of support vector machines and neural networks. The support vector machine is used to process high-dimensional features output from the hybrid neural network, and its sensitivity to boundary data is used to further improve the classification accuracy, thereby ensuring the ability to recognize and classify multiple targets in complex environments.

5. The multi-mode aerosonar buoy signal processing and target recognition method according to claim 1, characterized in that: The support vector machine adopts a nonlinear kernel function to adapt to the nonlinear relationship between different types of signals, thereby improving the classification effect of the system when processing multi-modal data. The use of this nonlinear kernel function can enhance the recognition accuracy of complex targets and cope with noise interference in the marine environment.

6. The multi-mode aerosonar buoy signal processing and target recognition method according to claim 2, characterized in that: The network structure of the convolutional neural network includes multiple convolutional layers and pooling layers, where the convolution operation is defined as: in, is the convolution output of the lth layer, is the input of the previous layer, is the convolution kernel weight of the lth layer, b (l) is the bias term. The pooling layer downsamples the convolution output to reduce the feature dimension and computational complexity. The pooling operation is defined as: in, Represents the pooling output of layer l, both maximum pooling and average pooling can be applied.

7. The multi-mode aerosonar buoy signal processing and target recognition method according to claim 3, characterized in that: The hybrid neural network structure adopts a combination of convolutional layer and recurrent neural network layer, wherein the recurrent neural network layer is used to capture the temporal characteristics of the signal, and the hidden layer state update formula of the recurrent neural network layer is: t =σ(W h h t-1 +W x x t +b h )in, h t is the hidden state at the tth moment, h t-1 is the hidden state at the previous moment, x t is the input at the current moment, W h and W x are weight matrices, b h is the bias term, and σ is the activation function.

Images