A method for identifying and analyzing the distribution characteristics of artificial fish reef piles

Abstract

The present invention relates to the technical field of marine ecological monitoring, and particularly relates to a method for identifying and analyzing the distribution characteristics of artificial reef piles, comprising the following steps: S1: Obtain echo signal data, including time delay, signal intensity, and frequency information; S2: Preprocess the echo signal data obtained in S1; S3: Based on the preprocessed echo signal data, identify the characteristic signals related to artificial reef piles; S4: Extract the distribution characteristics of artificial reef piles, including depth, shape, and distribution density; S5: Analyze and identify the distribution characteristics of different types of reef piles; S6: Generate a three-dimensional distribution map of the analysis results in S5 through visualization technology. In the present invention, by combining echo detection technology with the K-means clustering algorithm, the distribution characteristics of artificial reef piles can be efficiently identified and analyzed, and their distribution conditions can be intuitively displayed through three-dimensional visualization technology, thereby improving the accuracy of artificial reef pile monitoring.

Description

Technical Field

[0001] The present invention relates to the technical field of marine ecological monitoring, and particularly to a method for identifying and analyzing the distribution characteristics of artificial reef piles. Background Art

[0002] With the increasing attention to marine ecological protection and fishery resource management, the construction and monitoring of artificial reefs have become an important direction for marine environmental protection and the sustainable development of fishery resources; artificial reefs can provide habitats for marine organisms and promote the restoration and proliferation of fishery resources; traditional methods for monitoring the distribution of artificial reef piles mainly rely on manual diving, trawl sampling, etc. These methods are not only cumbersome and costly to operate, but also have low monitoring efficiency in large-scale sea areas and are difficult to achieve real-time and large-scale monitoring; in addition, with the increasing complexity of the marine environment, traditional methods have limited ability to accurately locate artificial reef piles and extract distribution characteristics in complex waters, and there is an urgent need for an efficient and accurate technical means to monitor and analyze marine artificial reef piles.

[0003] However, although existing echo detection technologies can be used for detecting the seabed topography, there are still many technical problems in the monitoring of artificial reef piles; firstly, most existing echo detection technologies adopt simple echo signal processing methods and cannot effectively distinguish different types of artificial reef piles and their distribution characteristics; secondly, in the monitoring of large-scale sea areas, how to accurately extract the depth, shape, distribution density and other characteristics of artificial reef piles and classify and identify them remains a technical problem. Summary of the Invention

[0004] Based on the above purposes, the present invention provides a method for identifying and analyzing the distribution characteristics of artificial reef piles.

[0005] A method for identifying and analyzing the distribution characteristics of artificial reef piles includes the following steps:

[0006] S1: Conduct echo detection on the target sea area through an echo detection device to obtain echo signal data, including time delay, signal intensity and frequency information;

[0007] S2: Preprocess the echo signal data obtained in S1, including signal denoising, filtering and time-domain to frequency-domain conversion;

[0008] S3: Based on the preprocessed echo signal data, identify the characteristic signals related to artificial reef piles;

[0009] S4: Extract the characteristics of the target signals detected in S3 to extract the distribution characteristics of artificial reef piles, including depth, shape and distribution density;

[0010] S5: Analyze the extracted distribution features using the K - means clustering algorithm to classify and identify the distribution features of different types of artificial reef piles;

[0011] S6: Generate a three - dimensional distribution map of the analysis results in S5 through visualization technology to display the distribution features of artificial reef piles.

[0012] Optionally, S1 specifically includes:

[0013] S11: Deploy echo detection equipment in the target sea area, including a sonar transmitter and a sonar receiver. The sonar transmitter is used to transmit acoustic wave signals to the seabed, and the sonar receiver is used to receive the reflected echo signals;

[0014] S12: Transmit acoustic wave signals through the sonar transmitter according to predetermined frequency and power parameters, where the frequency range is from 10 kHz to 500 kHz and the power range is from 1 W to 1000 W. The acoustic wave signals propagate in seawater and generate reflected echoes when encountering artificial reef piles;

[0015] S13: The sonar receiver receives the echo signals reflected from the artificial reef piles. The echo signals include time delay, signal intensity, and frequency information;

[0016] S14: Synchronize the received echo signal data with the time reference of the transmitted signal to accurately calculate the time delay of the echo signal and ensure the accuracy of signal intensity and frequency information.

[0017] Optionally, S14 specifically includes:

[0018] S141: Record the accurate time point of the transmitted signal when the sonar transmitter emits acoustic wave signals ;

[0019] S142: Record the time point of the received signal when the sonar receiver receives the echo signal ;

[0020] S143: Calculate the time delay of the echo signal according to the recorded transmission time and reception time : :

[0021] S144: Determine whether the echo signal is valid by comparing the calculated time delay with a preset time threshold . If , the echo signal is considered valid, and the corrected signal data is used for subsequent steps; otherwise, discard the echo signal data.

[0022] Optionally, S2 specifically includes:

[0023] S21: Apply the wavelet transform method to denoise the echo signal, specifically including applying the discrete wavelet transform to the echo signal, decomposing the signal into multiple frequency bands, and removing the noise components through the soft threshold denoising technique, only retaining the specified characteristic components of the signal;

[0024] S22: Apply a band-pass filter to the denoised echo signal, and set its cut-off frequency range to 10 kHz to 500 kHz to filter out the frequency components below 10 kHz and above 500 kHz;

[0025] S23: Perform a fast Fourier transform on the filtered echo signal to convert the time-domain signal into a frequency-domain signal and generate spectral data;

[0026] S24: Standardize the converted frequency-domain signal data, using the Z-score standardization method to normalize the signal intensity to the range with a mean of 0 and a standard deviation of 1 to ensure the comparability between different signals.

[0027] Optionally, the specific steps of S3 include:

[0028] S31: Use the preprocessed echo signal data to train a target detection model. This target detection model is constructed through a convolutional neural network, including an input layer, multiple convolutional layers, pooling layers, and fully connected layers, and is used to identify the characteristic signals related to artificial fish reef piles;

[0029] S32: Extract the spatial features and temporal features in the echo signal;

[0030] S33: Input the extracted features into the classification layer, and use the Softmax function to output the probability that the signal belongs to an artificial fish reef pile;

[0031] S34: Screen out the signals with probabilities higher than the threshold according to the preset probability threshold, and then determine that these signals correspond to artificial fish reef piles.

[0032] Optionally, the specific steps of S31 include:

[0033] S311: Label the preprocessed echo signal data as artificial fish reef piles and non-fish reef piles to form a labeled data set , where represents the i-th echo signal sample, represents the corresponding label, which is an artificial fish reef pile or a non-fish reef pile, and N is the number of echo signal samples;

[0034] S312: Divide the data set D into a training set and a validation set in a ratio of 80% for the training set and 20% for the validation set;

[0035] S313: Construct a convolutional neural network model, i.e., the target detection model structure, including an input layer, multiple convolutional layers, pooling layers, fully connected layers, and an output layer;

[0036] S314: Use the training set to train the target detection model; specifically including the following steps:

[0037] S3141: Pass the input data through the target detection model to calculate the predicted output ;

[0038] S3142: Use the cross-entropy loss function L to calculate the error between the predicted value and the true label;

[0039] S3143: Calculate the gradient of the model parameters according to the loss function L, and update the model parameters through an optimization algorithm to minimize the loss function;

[0040] S3144: Repeat the above S3141 - S3143 until the loss function converges or reaches the preset number of training epochs.

[0041] Optionally, the specific content of S4 includes:

[0042] S41: Use the time delay of the echo signal and the known sound wave propagation speed v, according to the formula: depth to calculate and extract the depth information of the artificial reef pile;

[0043] S42: Conduct morphological analysis on the detected echo signal, and use a morphological filter to process the signal image to extract the geometric shape features of the artificial reef pile, including the size, contour, and surface texture of the pile body;

[0044] S43: Based on the extracted depth and morphological features, apply spatial statistical analysis methods to calculate the distribution density of the artificial reef piles in the target sea area, specifically including counting the number of piles within a preset grid range and calculating the average density of the piles in each grid unit;

[0045] S44: Standardize the extracted depth, morphological, and distribution density features, using the Z-score standardization method to convert each feature into a standard normal distribution with a mean of 0 and a standard deviation of 1 to eliminate the influence of different feature dimensions on subsequent analysis.

[0046] Optionally, the specific content of S43 includes:

[0047] S431: Divide the target sea area into several grid units of equal size;

[0048] S432: Count the number of artificial fish reef piles in each grid cell, denoted as , where i represents the i-th grid cell;

[0049] S433: Calculate the area of each grid cell , with the unit of square meters;

[0050] S434: According to the number of piles and the area in each grid cell, calculate the distribution density in this grid cell. The formula is: , where is the distribution density of the i-th grid cell, with the unit of number of piles per square meter.

[0051] Optionally, the specific steps of S5 include:

[0052] S51: Randomly initialize the centroids of K clusters according to the pre-determined number of clusters , , , ;

[0053] S52: For each sample point , calculate its Euclidean distance from each centroid , and assign to the cluster to which the nearest clustering center belongs ;

[0054] S53: For each cluster , recalculate its clustering center as the mean of all sample points within the cluster;

[0055] S54: Repeat S52 - S53 until the positions of the clustering centers no longer change or reach the preset maximum number of iterations;

[0056] S55: According to the clustering results, classify the distribution characteristics of artificial fish reef piles into different distribution patterns, including high-density piles, low-density piles, and linear distribution piles.

[0057] Optionally, the specific steps of S6 include:

[0058] S61: Select a software system for three-dimensional data visualization, which includes geographic information system software or three-dimensional modeling software;

[0059] S62: Import the classification and recognition results and the corresponding distribution characteristic data obtained from the S5 data analysis into the visualization software system;

[0060] S63: Set corresponding visual parameter mappings according to the depth, shape, and distribution density characteristics of the artificial reef piles, specifically including:

[0061] Depth mapping: Map the depth data to the vertical coordinates in three-dimensional space to reflect the depth distribution of the reef piles.

[0062] Shape mapping: Convert the shape characteristics into the geometric shape parameters of the three-dimensional model.

[0063] Distribution density mapping: Map the distribution density data to visual attributes such as color, transparency, or size to distinguish different density regions.

[0064] S64: Based on the mapped feature data, use the modeling function of the visualization software to generate a three-dimensional distribution model of the artificial reef piles, specifically including integrating the position information, depth, and shape characteristics of each reef pile into the three-dimensional space by establishing a three-dimensional coordinate system, thereby generating a three-dimensional distribution map of the artificial reef piles.

[0065] Advantages of the present invention:

[0066] In the present invention, through the combination of echo detection technology and data analysis algorithms, an efficient and accurate method for identifying and analyzing the distribution characteristics of artificial reef piles is provided. By using echo detection equipment to obtain the echo signal data of artificial reef piles in the sea area and performing data preprocessing through precise signal processing steps, noise and interference can be effectively removed, and the accuracy of the data can be improved. By using a target detection model to identify the feature signals related to artificial reef piles and further extracting key features such as the depth, shape, and distribution density of the pile body, the efficiency and accuracy of artificial reef pile monitoring are improved. Compared with traditional manual diving and trawl sampling, real-time monitoring can be achieved and a wider sea area can be covered.

[0067] In the present invention, by using the K-means clustering algorithm to classify and analyze the extracted distribution characteristics, the distribution characteristics of different types of reef piles can be accurately identified, and the analysis results can be presented as a three-dimensional distribution map through visualization technology. This visualization method not only provides researchers with an intuitive distribution of reef piles, facilitates the monitoring and evaluation of different regions, but also helps decision-makers formulate more scientific fishery resource management and ecological protection measures. Description of the Drawings

[0068] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for use in the embodiments or the description of the prior art. Obviously, the following drawings are only those of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0069] Figure 1Schematic diagram of the method for identifying and analyzing the distribution characteristics of artificial reef piles according to the embodiments of the present invention;

[0070] Figure 2 Schematic diagram of the method for identifying characteristic signals related to artificial reef piles according to the embodiments of the present invention. Detailed implementation manners

[0071] The present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it should be noted here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments. For some well-known technologies, those skilled in the art can also adopt other alternative methods for implementation; moreover, the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0072] As Figure 1 - Figure 2 shown, a method for identifying and analyzing the distribution characteristics of artificial reef piles includes the following steps:

[0073] S1: Perform echo detection on the target sea area through an echo detection device to obtain echo signal data, including time delay, signal intensity, and frequency information;

[0074] S2: Preprocess the echo signal data obtained in S1, including signal denoising, filtering, and time-domain to frequency-domain conversion;

[0075] S3: Based on the preprocessed echo signal data, identify characteristic signals related to artificial reef piles;

[0076] S4: Extract the distribution characteristics of the artificial reef piles from the target signals detected in S3, including depth, shape, and distribution density;

[0077] S5: Use the K-means clustering algorithm to analyze the extracted distribution characteristics to classify and identify the distribution characteristics of different types of reef piles;

[0078] S6: Generate a three-dimensional distribution map through visualization technology for the analysis results in S5 to display the distribution characteristics of artificial reef piles.

[0079] S1 specifically includes:

[0080] S11: Deploy an echo detection device in the target sea area, including a sonar transmitter and a sonar receiver. The sonar transmitter is used to transmit acoustic wave signals to the seabed, and the sonar receiver is used to receive the reflected echo signals;

[0081] S12: Transmit acoustic wave signals through the sonar transmitter according to predetermined frequency and power parameters, where the frequency range is 10 kHz to 500 kHz, and the power range is 1 W to 1000 W. The acoustic wave signals propagate in seawater and generate reflected echoes when encountering artificial reef piles;

[0082] S13: The sonar receiver receives the echo signal reflected from the artificial reef pile. The echo signal includes time delay, signal strength, and frequency information.

[0083] S14: Synchronize the received echo signal data with the time reference of the transmitted signal to accurately calculate the time delay of the echo signal and ensure the accuracy of the signal strength and frequency information. Through the above steps, the echo detection device can accurately and real-time perform echo detection on the target sea area, obtain high-quality echo signal data, and provide accurate signal input for the subsequent data preprocessing steps.

[0084] The specific synchronization process in S14 includes:

[0085] S141: When the sonar transmitter emits a sound wave signal, record the accurate time point of the transmitted signal ;

[0086] S142: When the sonar receiver receives the echo signal, record the time point of the received signal ;

[0087] S143: According to the recorded transmission time and reception time , calculate the time delay of the echo signal ; The calculation formula is: :

[0088] S144: By comparing the calculated time delay with the preset time threshold , judge whether the echo signal is valid. If , it is considered that the echo signal is valid, and the corrected signal data is used for the subsequent steps; otherwise, discard the echo signal data. Through the above steps, the method realizes the precise synchronization process of the echo signal data with the time reference of the transmitted signal, ensures the accurate calculation of the time delay, signal strength, and frequency information, and provides reliable data support for the precise identification and analysis of the distribution characteristics of artificial reef piles.

[0089] S2 specifically includes:

[0090] S21: Use the wavelet transform method to denoise the echo signal, specifically including applying the discrete wavelet transform (DWT) to the echo signal, decomposing the signal into multiple frequency bands, and removing the noise components through the soft threshold denoising technique, only retaining the specified characteristic components of the signal, thereby improving the signal-to-noise ratio of the signal.

[0091] S22: Apply a band - pass filter to the denoised echo signal, with its cut - off frequency range set from 10 kHz to 500 kHz, to filter out frequency components below 10 kHz and above 500 kHz, ensure that the signal frequency range meets the requirements of subsequent analysis steps, and effectively remove environmental noise interference;

[0092] S23: Perform a Fast Fourier Transform (FFT) on the filtered echo signal to convert the time - domain signal into a frequency - domain signal and generate spectral data for facilitating the analysis and processing of the signal frequency characteristics in subsequent target detection and feature extraction steps;

[0093] S24: Standardize the converted frequency - domain signal data using the Z - score standardization method to normalize the signal intensity within a range with a mean of 0 and a standard deviation of 1, ensuring the comparability between different signals; Through the above steps, the data pre - processing process can effectively remove noise and unnecessary frequency components in the echo signal, ensuring that the data used in subsequent target detection and feature extraction steps has high quality and high accuracy.

[0094] S3 specifically includes:

[0095] S31: Use the pre - processed echo signal data to train a target detection model, which is constructed by a convolutional neural network and includes an input layer, multiple convolutional layers, pooling layers, and fully - connected layers for identifying feature signals related to artificial reef piles;

[0096] S32: Extract spatial features and temporal features from the echo signal;

[0097] S33: Input the extracted features into the classification layer and use the Softmax function to output the probability that the signal belongs to an artificial reef pile;

[0098] S34: Screen out signals with probabilities higher than the threshold according to a preset probability threshold, and then determine that these signals correspond to artificial reef piles; Through the above steps, the echo signal features related to artificial reef piles can be effectively identified, improving the accuracy and reliability of the identification and ensuring the data quality relied on in subsequent feature extraction and data analysis steps.

[0099] S31 specifically includes:

[0100] S311: Label the pre - processed echo signal data as artificial reef piles and non - reef piles to form a labeled data set , where represents the i - th echo signal sample, represents the corresponding label, which is an artificial reef pile or a non - reef pile, and N is the number of echo signal samples;

[0101] S312: Divide the dataset D into a training set and a validation set where the training set is used for training the object detection model, and the validation set is used for performance evaluation and hyperparameter tuning of the object detection model. The splitting ratio is 80% for the training set and 20% for the validation set;

[0102] S313: Build a convolutional neural network model, i.e., the object detection model structure, including an input layer, multiple convolutional layers, pooling layers, fully connected layers, and an output layer; The specific structure includes:

[0103] Input layer: Used to receive the preprocessed echo signal data, with a shape of (H, W, C), where H is the height, W is the width, and C is the number of channels;

[0104] Convolutional layer: Used to apply multiple convolutional kernels for feature extraction and output feature maps;

[0105] Pooling layer: Used to reduce the dimension of the feature maps and improve the generalization ability of the model by using the max pooling or average pooling method;

[0106] Fully connected layer: Used to expand the pooled feature maps, connect to fully connected neurons, and perform high-level feature combination;

[0107] Output layer: Used to output the classification probability distribution by using the Softmax activation function .

[0108] S314: Use the training set to train the object detection model; Specifically, it includes the following steps:

[0109] S3141: Pass the input data through the object detection model and calculate the predicted output ;

[0110] S3142: Use the cross-entropy loss function L to calculate the error between the predicted value and the true label. The expression of the cross-entropy loss function L is: , where L represents the total loss of the model; represents the total number of samples (or the size of the batch); represents the true label (target value) of the i-th sample. For a binary classification task, takes values of 0 or 1; represents the predicted probability of the i-th sample, that is, the probability that the model predicts the sample i belongs to class 1; is the natural logarithm of the probability that the i-th sample is predicted to belong to class 1; is the natural logarithm of the probability that the i-th sample is predicted to belong to class 0; the cross-entropy loss measures the difference between the probability distribution output by the model and the actual labels; if the probability predicted by the model matches the true label, the loss is small; if the prediction result deviates greatly from the actual label, the loss is large; When , the loss term is , that is, the closer the model is to the true label 1, the smaller the loss; if the probability predicted by the model is very small (i.e., close to 0), the loss will be very large; When

[0111] S3143: Calculate the gradient of the model parameters according to the loss function L, and update the model parameters through an optimization algorithm (such as Stochastic Gradient Descent SGD or Adam optimizer) to minimize the loss function;

[0112] S3144: Repeat the above S3141 - S3143 until the loss function converges or reaches the preset number of training epochs;

[0113] S315: Use the validation set to evaluate the performance of the trained object detection model, calculate the metrics of accuracy, recall, and F1-score to ensure that the model has good generalization ability and recognition accuracy; through the above steps, the convolutional neural network object detection model is trained using the preprocessed echo signal data with annotations, ensuring that the model can accurately identify the echo signal characteristics related to artificial reef piles, improving the accuracy and reliability of object detection.

[0114] S4 specifically includes:

[0115] S41: Utilize the time delay of the echo signal and the known sound wave propagation speed v (usually taken as 1500 m / s in seawater), and according to the formula: depth , calculate and extract the depth information of the artificial reef pile to ensure the accuracy and reliability of depth measurement;

[0116] S42: Conduct morphological analysis on the detected echo signal, and use morphological filters (such as opening operation and closing operation) to process the signal image to extract the geometric shape characteristics of the artificial reef pile, including the size, contour, and surface texture of the pile body;

[0117] S43: Based on the extracted depth and morphological features, apply spatial statistical analysis methods to calculate the distribution density of artificial reef piles in the target sea area, specifically including counting the number of pile bodies within a preset grid range and calculating the average density of pile bodies in each grid cell;

[0118] S44: Standardize the extracted depth, morphology, and distribution density features. Using the Z-score standardization method, convert each feature into a standard normal distribution with a mean of 0 and a standard deviation of 1 to eliminate the influence of different feature dimensions on subsequent analysis and ensure the consistency and comparability of feature data. Through the above steps, comprehensive feature extraction of the target signal detected in S3 is achieved, and key distribution features such as the depth, morphology, and distribution density of artificial reef piles are accurately obtained, providing high-quality feature data support for subsequent data analysis and visualization.

[0119] S43 specifically includes:

[0120] S431: Divide the target sea area into several equal-sized grid cells. The size of the grid cells is preset according to the specific scope and resolution requirements of the target sea area to ensure the accuracy of distribution density calculation.

[0121] S432: Count the number of artificial reef piles in each grid cell, denoted as , where i represents the i-th grid cell.

[0122] S433: Calculate the area of each grid cell, with the unit of square meters.

[0123] S434: According to the number of piles and area in each grid cell, calculate the distribution density in this grid cell. The formula is: , where is the distribution density of the i-th grid cell, with the unit of number of piles per square meter. Through the above steps, accurate calculation of the distribution density of artificial reef piles in the target sea area is achieved, providing an accurate and comparable data basis for the subsequent clustering analysis step.

[0124] S5 specifically includes:

[0125] S51: Randomly initialize the centroids , , , of K clusters according to the pre-determined number of clusters.

[0126] S52: For each sample point , calculate its Euclidean distance from each centroid , and assign to the cluster of the nearest clustering center. The formula is:

[0127] , where M is the feature dimension. is the j-th eigenvalue of the i-th sample, is the j-th eigenvalue of the k-th cluster centroid;

[0128] S53: For each cluster , recalculate its clustering center is the mean of all sample points within the cluster, and the formula is: ;

[0129] S54: Repeat S52 - S53 until the position of the clustering center no longer changes or reaches the preset maximum number of iterations;

[0130] S55: According to the clustering results, classify the distribution characteristics of artificial reef piles into different distribution patterns, including high - density piles, low - density piles, and linear - distribution piles; Through the above steps, the K - means clustering algorithm can effectively classify and identify the extracted distribution characteristics of artificial reef piles, accurately distinguish different types of reef - pile distribution patterns, and improve the accuracy and reliability of distribution - characteristic analysis.

[0131] S6 specifically includes:

[0132] S61: Select a software system for three - dimensional data visualization. The software system includes Geographic Information System (GIS) software or three - dimensional modeling software to support the generation of high - precision three - dimensional distribution maps;

[0133] S62: Import the classification and recognition results and the corresponding distribution - characteristic data (depth, shape, distribution density) obtained in S5 data analysis into the visualization software system, ensuring that the data format matches the requirements of the software system;

[0134] S63: According to the depth, shape, and distribution - density characteristics of artificial reef piles, set the corresponding visualization - parameter mappings, specifically including:

[0135] Depth mapping: Map the depth data to the vertical coordinate in three - dimensional space to reflect the depth distribution of reef piles;

[0136] Shape mapping: Convert the shape characteristics into geometric - shape parameters of the three - dimensional model, such as size, contour, and surface texture;

[0137] Distribution - density mapping: Map the distribution - density data to visual attributes such as color, transparency, or size to distinguish different density regions;

[0138] S64: Based on the mapped feature data, use the modeling function of visualization software to generate a three-dimensional distribution model of artificial reef piles. Specifically, by establishing a three-dimensional coordinate system, integrate the position information, depth, and morphological characteristics of each reef pile into three-dimensional space, thereby generating a three-dimensional distribution map of artificial reef piles; through the above steps, the spatial distribution, morphology, and density characteristics of artificial reef piles can be accurately displayed in three dimensions, providing an intuitive and efficient decision-making support tool for subsequent research and applications.

[0139] The above are only the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention.

Claims

1. A method for identifying and analyzing the distribution characteristics of artificial reefs, characterized in that: The following steps are involved: S1: Use echo detection equipment to perform echo detection on the target sea area and obtain echo signal data, including time delay, signal strength and frequency information; S2: preprocessing the echo signal data obtained by S1, including signal denoising, filtering and time domain-frequency domain conversion; S3: Based on the pre-processed echo signal data, identifying characteristic signals related to the artificial reef pile; S4: Extract the features of the target signal detected by S3 and extract the distribution features of the artificial reef pile, including depth, shape and distribution density, including: S41: Using the time delay of the echo signal and the known sound wave propagation speed v, according to the formula: depth , calculate and extract the depth information of artificial reef piles; S42: performing morphological analysis on the detected echo signal, and processing the signal image with a morphological filter to extract geometric shape features of the artificial fish reef pile, including the size, contour and surface texture of the pile; S43: Based on the extracted depth and morphological features, a spatial statistical analysis method is applied to calculate the distribution density of artificial reef piles in the target sea area, specifically including counting the number of piles within a preset grid range and calculating the average density of the piles in each grid unit; S44: Standardize the extracted depth, morphology and distribution density features. Use the Z-score standardization method to convert each feature into a standard normal distribution with a mean of 0 and a standard deviation of 1 to eliminate the influence of different feature dimensions on subsequent analysis. The S43 specifically includes: S431: Divide the target sea area into a number of grid cells of equal size; S432: Count the number of artificial reefs in each grid cell, recorded as , where i represents the i-th grid unit; S433: Calculate the area of ​​each grid cell , the unit is square meters; S434: Based on the number of piles in each grid unit and area , calculate the distribution density within the grid unit, the formula is: ,in, is the distribution density of the ith grid unit, in units of piles per square meter; S5: Use K-means clustering algorithm to analyze the extracted distribution features to classify and identify the distribution characteristics of different types of fish reef piles; S6: Generate a three-dimensional distribution map using visualization technology based on the analysis results in S5 to show the distribution characteristics of the artificial reef piles.

2. The method for identifying and analyzing the distribution characteristics of artificial reefs according to claim 1, characterized in that: The S1 specifically includes: S11: deploying echo detection equipment in the target sea area, including a sonar transmitter and a sonar receiver, wherein the sonar transmitter is used to transmit a sound wave signal to the seabed, and the sonar receiver is used to receive the reflected echo signal; S12: transmitting a sound wave signal through a sonar transmitter according to predetermined frequency and power parameters, wherein the frequency range is 10kHz to 500kHz, and the power range is 1W to 1000W, and the sound wave signal propagates in the seawater and generates a reflected echo when encountering an artificial reef pile; S13: The sonar receiver receives the echo signal reflected from the artificial reef pile, wherein the echo signal includes time delay, signal strength and frequency information; S14: Synchronize the received echo signal data with the time reference of the transmitted signal to accurately calculate the time delay of the echo signal and ensure the accuracy of the signal strength and frequency information.

3. The method for identifying and analyzing the distribution characteristics of artificial reefs according to claim 2, characterized in that: The S14 specifically includes: S141: When the sonar transmitter emits a sound wave signal, record the exact time point of the signal emission ; S142: When the sonar receiver receives the echo signal, the time point of receiving the signal is recorded ; S143: According to the recorded launch time and receiving time , calculate the time delay of the echo signal : S144: Time delay calculated by comparison With the preset time threshold , to determine whether the echo signal is valid, if , the echo signal is considered valid and the corrected signal data is used in subsequent steps; otherwise, the echo signal data is discarded.

4. The method for identifying and analyzing the distribution characteristics of artificial reefs according to claim 1, characterized in that: The S2 specifically includes: S21: performing denoising processing on the echo signal by using a wavelet transform method, specifically including applying discrete wavelet transform to the echo signal, decomposing the signal into multiple frequency bands, and removing noise components by using a soft threshold denoising technique, and only retaining designated characteristic components of the signal; S22: applying a bandpass filter to the denoised echo signal, with a cutoff frequency range set to 10 kHz to 500 kHz, so as to filter out frequency components below 10 kHz and above 500 kHz; S23: performing fast Fourier transform on the filtered echo signal to convert the time domain signal into a frequency domain signal to generate spectrum data; S24: The converted frequency domain signal data is standardized by using a Z-score standardization method to normalize the signal strength to a range of a mean of 0 and a standard deviation of 1 to ensure comparability between different signals.

5. The method for identifying and analyzing the distribution characteristics of artificial reefs according to claim 1, characterized in that: The S3 specifically includes: S31: using the preprocessed echo signal data to train a target detection model, the target detection model is constructed by a convolutional neural network, including an input layer, multiple convolutional layers, a pooling layer and a fully connected layer, and is used to identify characteristic signals related to artificial reefs; S32: extracting spatial features and temporal features from the echo signal; S33: Input the extracted features into the classification layer, and use the Softmax function to output the probability that the signal belongs to the artificial fish reef pile; S34: Filter out signals with probabilities higher than the threshold according to a preset probability threshold, and then determine that these signals correspond to artificial fish reef piles.

6. The method for identifying and analyzing the distribution characteristics of artificial reefs according to claim 5, characterized in that: The S31 specifically includes: S311: Label the preprocessed echo signal data according to artificial fish reef piles and non-fish reef piles to form a labeled data set ,in represents the i-th echo signal sample, Indicates the corresponding label, which is an artificial fish reef or a non-fish reef, and N is the number of echo signal samples; S312: Divide the data set D into training sets in proportion and validation set , the split ratio is 80% training set and 20% validation set; S313: Construct a convolutional neural network model, i.e., a target detection model structure, including an input layer, multiple convolutional layers, a pooling layer, a fully connected layer, and an output layer; S314: Using the training set Train the target detection model; specifically include the following steps: S3141: Input data Through the target detection model, calculate the predicted output ; S3142: Use the cross entropy loss function L to calculate the error between the predicted value and the true label; S3143: performing gradient calculation on the model parameters according to the loss function L, and updating the model parameters through an optimization algorithm to minimize the loss function; S3144: Repeat the above S3141-S3143 until the loss function converges or reaches a preset number of training rounds.

7. The method for identifying and analyzing the distribution characteristics of artificial reefs according to claim 1, characterized in that: The S5 specifically includes: S51: Randomly initialize the centroids of K clusters according to the predetermined number of clusters , , , ; S52: For each sample point , calculate its relative position to each centroid Euclidean distance , and Assign to the cluster with the closest cluster center ; S53: For each cluster , recalculate its cluster center is the mean of all sample points in the cluster; S54: Repeat S52-S53 until the cluster center position does not change or the preset maximum number of iterations is reached; S55: Based on the clustering results, the distribution characteristics of artificial reef piles are classified into different distribution patterns, including high-density piles, low-density piles and linear distribution piles.

8. The method for identifying and analyzing the distribution characteristics of artificial reefs according to claim 1, characterized in that: The S6 specifically includes: S61: Selecting a software system for three-dimensional data visualization, wherein the software system includes geographic information system software or three-dimensional modeling software; S62: importing the classification recognition results and corresponding distribution feature data obtained in the data analysis of S5 into the visualization software system; S63: According to the depth, shape and distribution density characteristics of the artificial reef, corresponding visualization parameter mapping is set, including: Depth mapping: Mapping depth data into vertical coordinates in three-dimensional space to reflect the depth distribution of fish reef piles; Morphological mapping: converting morphological features into geometric shape parameters of a three-dimensional model; Distribution density mapping: Map distribution density data to visual attributes such as color, transparency, or size to distinguish areas of different densities; S64: Based on the mapped feature data, a three-dimensional distribution model of the artificial fish reef pile is generated using the modeling function of the visualization software, specifically including integrating the location information, depth and morphological characteristics of each fish reef pile into the three-dimensional space by establishing a three-dimensional coordinate system, thereby generating a three-dimensional distribution map of the artificial fish reef pile.

Images