Sonar imaging optimization method for underwater environment condition monitoring

Abstract

The invention discloses a sonar imaging optimization method for monitoring an underwater environment condition, and relates to the technical field of underwater surveying and mapping, sonar matrixes constructed by a plurality of sonar front ends are utilized to simultaneously emit sound wave signals with different frequencies, so that the same feature can be determined to be correspondingly identified under different frequencies, and the accuracy of the sonar imaging optimization is improved. Corresponding underwater feature models are generated according to recognition results of all the sonar front ends, distribution areas of the underwater feature models are divided according to distribution conditions of the underwater feature models, clustering evaluation is carried out, the divided areas are optimized according to clustering evaluation results, and the underwater feature models are obtained. According to the method, the underwater feature models of the same feature identified by the plurality of sonar front ends and at different positions are clustered, and a unique underwater factor label is obtained, so that the constructed underwater simulation image is correspondingly optimized, and the situation that a plurality of underwater feature models exist in the underwater simulation image of the same feature is avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of underwater surveying and mapping, and specifically to a sonar imaging optimization method for monitoring the underwater environmental conditions. Background Art

[0002] Sonar is currently the most effective device for humans to detect underwater targets. Image sonar is a sonar that can directly image acoustic detection information. In many fields such as ocean development, ocean scientific research, and underwater security, image sonar plays an indispensable role;

[0003] Using multiple sonar front-ends to construct a sonar matrix to obtain sonar signal data in a larger sea area is a commonly used technical means at present. When using multiple sonar front-ends for detection, different sonar front-ends often produce different results for the same target in the sea area, which may lead to distorted detection results and mutual interference. How to effectively cluster the detection results of different sonar front-ends to eliminate the mutual interference between different sonar front-ends is the problem we need to solve. For this reason, a sonar imaging optimization method for monitoring the underwater environmental conditions is provided. Summary of the Invention

[0004] The purpose of the present invention is to provide a sonar imaging optimization method for monitoring the underwater environmental conditions.

[0005] The purpose of the present invention can be achieved by the following technical solutions: A sonar imaging optimization method for monitoring the underwater environmental conditions includes the following steps:

[0006] Step S1: Arrange sonar front-ends at different positions in the target ocean area, and construct a corresponding sonar front-end matrix according to the arranged sonar front-ends;

[0007] Step S2: Transmit corresponding acoustic wave signals to the target ocean area through the sonar front-end matrix, and receive the corresponding reflected signals;

[0008] Step S3: Process the reflected signals received by the sonar front-end matrix to complete the construction of an underwater simulation image of the target ocean area;

[0009] Step S4: Optimize the constructed underwater simulation image in regions, and generate corresponding underwater factor labels in the underwater simulation image according to the optimization results.

[0010] Further, the process of arranging sonar front-ends at different positions in the target ocean area and constructing a corresponding sonar front-end matrix includes:

[0011] Construct a three-dimensional coordinate system, select the target ocean area, generate an ocean plane model according to the range of the target ocean area, and map the ocean plane model into the three-dimensional coordinate system;

[0012] Set a number of sonar front-ends within the target ocean area;

[0013] Select the location of any sonar front-end as the origin and map it into the three-dimensional coordinate system, and map the locations of the other set sonar front-ends into the three-dimensional coordinate system;

[0014] Obtain the range of the acoustic wave signal emission angle corresponding to each sonar front-end, and generate an acoustic wave signal coverage area in the three-dimensional coordinate system according to the range of the acoustic wave signal emission angle;

[0015] According to the number of sonar front-ends, construct a sonar front-end matrix composed of a number of blank matrix units, and associate each sonar front-end with a blank matrix unit;

[0016] Import the location coordinates of the sonar front-ends into the corresponding blank matrix units.

[0017] Further, the process of transmitting the corresponding acoustic wave signals to the target ocean area through the sonar front-end matrix and receiving the corresponding reflected signals includes:

[0018] The sonar front-end emits acoustic wave signals of different frequencies according to the corresponding range of the acoustic wave signal emission angle, and each sonar front-end emits acoustic wave signals in a rotation manner. After all sonar front-ends complete the rotation, it is recorded as a rotation period, and the reflected signals corresponding to each acoustic wave signal within the rotation period are collected;

[0019] Record the signal emission time and signal parameters of the acoustic wave signals emitted by each acoustic wave signal;

[0020] Summarize the acoustic wave signals and reflected signals within a rotation period to obtain the corresponding signal data set.

[0021] Further, the process of processing the reflected signals received by the sonar front-end matrix includes:

[0022] Construct a two-dimensional coordinate system of time versus signal parameters corresponding to each sonar front-end;

[0023] Set the corresponding frequency fluctuation range according to the frequency table;

[0024] Filter the reflected signals received by each sonar front-end through the set frequency fluctuation range;

[0025] Map the filtered reflected signals into the corresponding coordinate system to obtain a spectrogram corresponding to each sonar terminal;

[0026] Input each of the obtained spectrograms into the trained convolutional neural network model, identify the spectral features in each spectrogram through the convolutional neural network model, and convert the identified spectral features into corresponding underwater factor features.

[0027] Further, the training process of the convolutional neural network model is as follows:

[0028] Collect sample data, where the sample data is the spectral signal corresponding to the reflected signal received after the sonar terminal emits a sound wave signal according to the frequencies in the frequency table;

[0029] Associate each sample data with the corresponding true feature, and input the collected sample data into the convolutional neural network model;

[0030] The convolutional neural network model outputs the corresponding predicted features according to the input sample data;

[0031] Input the output predicted features and true features into the loss function to obtain the feature loss value;

[0032] Repeat the above operations until the set specific number of training times or the feature loss value obtained by the loss function is reached, thus completing the training process of the convolutional neural network model.

[0033] Further, the construction process of the underwater simulation image of the target ocean area includes:

[0034] Successively select the sonar front end as the reference end, and according to the signal frequency of the sound wave signal emitted by the reference end, obtain the spectrograms of the reflected signals of each sonar front end corresponding to this signal frequency;

[0035] Record the time corresponding to the identified spectral features as t j ;

[0036] Record the sound wave signal emission time of the sonar terminal as t0;

[0037] Then obtain the distance from the sonar terminal to the underwater factor feature corresponding to this spectral feature;

[0038] Select the moment corresponding to this spectral feature of any other sonar terminal, and further obtain the azimuth of the underwater factor feature corresponding to this spectral feature;

[0039] According to the obtained azimuth and distance of this underwater factor feature, generate the corresponding underwater feature model in the three-dimensional coordinate system;

[0040] Complete the positioning of all spectral features corresponding to underwater factor features in the spectrogram, and generate the corresponding underwater feature model.

[0041] Further, the process of optimizing the constructed underwater simulation image region by region and generating corresponding underwater factor labels within the underwater simulation image according to the optimization results includes:

[0042] Mark the spatial coordinates of the underwater feature models corresponding to each sonar front end as the reference end in the three-dimensional coordinate system;

[0043] Obtain the two spatial coordinates with the largest marked straight-line distance, and generate a corresponding connection line between the two spatial coordinates;

[0044] Generate a corresponding spherical space with the connection line between the two connected spatial coordinates as the diameter, and mark the generated spherical space as the corresponding feature partition;

[0045] Obtain the center point coordinates of the spherical space, and obtain the straight-line distances from the center point to each spatial coordinate within the spherical space;

[0046] Furthermore, obtain the clustering coefficient Jx of this feature partition;

[0047] If Jx≥0, it means that the clustering degree of this feature partition is high. Then, take the center point of this feature partition as the clustering point of the underwater feature model, and cluster the spatial coordinates of all underwater feature models within the feature partition to the clustering point, thereby completing the optimization of this feature partition and generating an underwater factor label corresponding to this clustering point;

[0048] When Jx<0, it means that the clustering degree of this feature partition is low. Then, eliminate the two connected spatial coordinates, and select the two spatial coordinates with the largest straight-line distance from the remaining spatial coordinates to generate a new spherical space as the feature partition, and then re-obtain the clustering coefficient corresponding to this feature partition, and so on, until the optimization of the feature partition is completed.

[0049] Compared with the prior art, the beneficial effects of the present invention are:

[0050] The sonar matrix constructed by using multiple sonar front ends simultaneously emits acoustic wave signals of different frequencies, enabling the determination that the same feature can be correspondingly recognized at different frequencies. Then, generate corresponding underwater feature models according to the recognition results of each sonar front end, and according to the distribution of the underwater feature models, divide the distribution area of the underwater feature models into regions, and perform clustering evaluation. Optimize the divided regions according to the clustering evaluation results, thereby clustering the underwater feature models of the same feature recognized by multiple sonar front ends and located at different positions, obtaining a unique underwater factor label, so that the constructed underwater simulation image is correspondingly optimized, and avoiding the situation where there are multiple underwater feature models of the same feature in the underwater simulation image. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the accompanying drawings required for use in the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other accompanying drawings can also be obtained based on these drawings.

[0052] Figure 1 It is a flowchart of the present invention. Specific implementation manners

[0053] As Figure 1 shown, a sonar imaging optimization method for underwater environmental condition monitoring includes the following steps:

[0054] Step S1: Arrange the sonar front ends at different positions in the target ocean area, and construct a corresponding sonar front end matrix according to the arranged sonar front ends;

[0055] Step S2: Transmit corresponding acoustic wave signals to the target ocean area through the sonar front end matrix, and receive the corresponding reflected signals;

[0056] Step S3: Process the reflected signals received by the sonar front end matrix to complete the construction of an underwater simulation image of the target ocean area;

[0057] Step S4: Perform regional optimization on the constructed underwater simulation image, and generate corresponding underwater factor labels in the underwater simulation image according to the optimization results.

[0058] It should be further noted that in the specific implementation process, the process of arranging the sonar front ends at different positions in the target ocean area and constructing a corresponding sonar front end matrix according to the arranged sonar front ends includes:

[0059] Construct a three-dimensional coordinate system of X, Y, and Z, select the target ocean area, generate an ocean plane model according to the range of the target ocean area, and map the ocean plane model into the three-dimensional coordinate system;

[0060] Set a number of sonar front ends in the target ocean area;

[0061] Select the location of any sonar front end as the origin and map it into the three-dimensional coordinate system, and map the locations of the other arranged sonar front ends into the three-dimensional coordinate system;

[0062] Obtain the acoustic wave signal emission angle range corresponding to each sonar front end, and generate an acoustic wave signal coverage area in the three-dimensional coordinate system according to the acoustic wave signal emission angle range;

[0063] Among them, the superposition of the acoustic wave signal coverage areas of all sonar front ends can cover the entire target ocean area;

[0064] Construct a sonar front-end matrix composed of several blank matrix units according to the number of sonar front-ends, and associate each sonar front-end with a blank matrix unit respectively;

[0065] Import the location coordinates of the sonar front-end into the corresponding blank matrix unit.

[0066] It should be further noted that in the specific implementation process, the process of emitting corresponding acoustic wave signals to the target ocean area through the sonar front-end matrix and receiving the corresponding reflected signals includes:

[0067] Number each sonar front-end, denoted as i, where i = 1, 2,..., n;

[0068] The sonar front-end emits acoustic wave signals of different frequencies according to the corresponding acoustic wave signal emission angle range, and each sonar front-end emits acoustic wave signals in a rotational manner. After completing the rotation of all sonar front-ends, it is recorded as a rotation period, and the reflected signals corresponding to each acoustic wave signal within the rotation period are collected;

[0069] Record the signal emission time and signal parameters of the acoustic wave signals emitted by each acoustic wave signal. The signal parameters include signal frequency and signal bandwidth;

[0070] Summarize the acoustic wave signals and reflected signals within a rotation period to obtain the corresponding signal dataset.

[0071] It should be further noted that in the specific implementation process, the rotational emission of acoustic wave signals is specifically as follows:

[0072] Set different frequencies equal to the number of sonar front-ends set, and generate a corresponding frequency table;

[0073] Assign a corresponding and distinct frequency to each sonar front-end as the initial frequency;

[0074] Each sonar front-end emits the corresponding acoustic wave signal according to the assigned initial frequency;

[0075] After completing the emission of the acoustic wave signals, re-assign new frequencies different from the initial frequencies to each sonar front-end, and re-emit the acoustic wave signals until the rotation of all different frequencies in the frequency table is completed;

[0076] Illustrate with an example:

[0077] Set sonar front-ends A, B, and C;

[0078] Set frequencies a, b, and c;

[0079] Then assign the initial frequency a to sonar front-end A, the initial frequency b to sonar front-end B, and the initial frequency c to sonar front-end C;

[0080] After the acoustic wave signal is transmitted, the frequency of sonar front end A is reassigned to b, the frequency of sonar front end B is reassigned to c, and the frequency of sonar front end C is reassigned to a;

[0081] After the acoustic wave signal is transmitted, the frequency of sonar front end A is reassigned to c, the frequency of sonar front end B is reassigned to a, and the frequency of sonar front end C is reassigned to b;

[0082] Thus, the transmission of the acoustic wave signal for one rotation cycle is completed.

[0083] It should be further noted that in the specific implementation process, the process of processing the reflected signals received by the sonar front end matrix includes:

[0084] Construct a two-dimensional coordinate system of time versus signal parameters corresponding to each sonar front end;

[0085] Set the corresponding frequency fluctuation range according to the frequency table;

[0086] Filter the reflected signals received by each sonar front end through the set frequency fluctuation range;

[0087] Map the filtered reflected signals into the corresponding coordinate system to obtain spectrograms corresponding to each sonar terminal;

[0088] When receiving sonar reflected signals, a lot of interference signals will also be received. By setting the frequency fluctuation range, most of the interference signals can be filtered, thereby improving the efficiency for the subsequent identification and extraction of spectral features;

[0089] Input the obtained spectrograms into the trained convolutional neural network model, identify the spectral features in each spectrogram through the convolutional neural network model, and convert the identified spectral features into corresponding underwater factor features.

[0090] It should be further noted that in the specific implementation process, the training process of the convolutional neural network model is specifically as follows:

[0091] Collect sample data, where the sample data is the spectral signal corresponding to the reflected signal received by the sonar terminal after transmitting the acoustic wave signal according to the frequency in the frequency table;

[0092] Associate the corresponding true features with each sample data, and input the collected sample data into the convolutional neural network model;

[0093] The convolutional neural network model outputs the corresponding predicted features according to the input sample data;

[0094] Input the output predicted features and true features into the loss function to obtain the feature loss value;

[0095] Repeat the above operations until the set number of specific training times or the feature loss value obtained by the loss function is reached, thus completing the training process of the convolutional neural network model.

[0096] It should be further noted that in the specific implementation process, the construction process of the underwater simulation image of the target ocean area includes:

[0097] Select the sonar front end with label i = 1 as the reference end, and according to the signal frequency of the acoustic wave signal emitted by the reference end, obtain the spectrogram of the reflected signal of each sonar front end corresponding to this signal frequency;

[0098] Label the spectrogram features identified in each spectrogram in sequence, denoted as j, where j = 1, 2,..., m;

[0099] Record the time corresponding to the identified spectrogram feature as t j ;

[0100] Record the acoustic wave signal emission time of the sonar terminal as t0;

[0101] Then obtain the distance from the sonar terminal to the underwater factor feature corresponding to this spectrogram feature, denoted as d j , where:

[0102]

[0103] where c is the propagation speed of acoustic waves in seawater, and Δt j is the time interval from t0 to t j ;

[0104] Select the moment corresponding to this spectrogram feature of any other sonar terminal, denoted as t dui ;

[0105] Then obtain the azimuth of the underwater factor feature corresponding to this spectrogram feature, denoted as θ, and it satisfies:

[0106]

[0107] where, Δt dui is the time interval from t0 to t dui , and b is the distance between the sonar front end emitting the acoustic wave signal and the selected sonar front end;

[0108] Generate a corresponding underwater feature model in the three-dimensional coordinate system according to the obtained azimuth and distance of the underwater factor feature;

[0109] Complete the positioning of all spectrogram features corresponding to underwater factor features in the spectrogram, and generate corresponding underwater feature models;

[0110] Then, select the sonar front end with label i = 2 as the reference end, and so on. Generate the corresponding underwater feature model in the three-dimensional coordinate system. After completing the construction of all underwater feature models, complete the underwater simulation image corresponding to the target ocean area.

[0111] It should be further noted that in the specific implementation process, the process of dividing the constructed underwater simulation image into regions for optimization and generating the corresponding underwater factor labels in the underwater simulation image according to the optimization results includes:

[0112] Mark the spatial coordinates of the underwater feature models corresponding to each sonar front end as the reference end in the three-dimensional coordinate system;

[0113] Obtain the two spatial coordinates with the largest marked straight-line distance, and generate a corresponding connection line between the two spatial coordinates;

[0114] Generate a corresponding spherical space with the connection line of the two connected spatial coordinates as the diameter, and mark the generated spherical space as the corresponding feature partition;

[0115] Obtain the center point coordinates of the spherical space, and obtain the straight-line distance from the center point to each spatial coordinate in the spherical space, and then record the corresponding straight-line distance as L i ;

[0116] Then obtain the clustering coefficient of this feature partition, denoted as Jx, where:

[0117]

[0118] Among them, ω is the standard clustering value;

[0119] If Jx ≥ 0, it means that the clustering degree of this feature partition is high. Then, take the center point of this feature partition as the clustering point of this underwater feature model, and cluster the spatial coordinates of all underwater feature models in the feature partition to the clustering point, thereby completing the optimization of this feature partition and generating the underwater factor label corresponding to this clustering point;

[0120] When Jx < 0, it means that the clustering degree of this feature partition is low. Then, eliminate the two connected spatial coordinates, and select the two spatial coordinates with the largest straight-line distance from the remaining spatial coordinates to generate a new spherical space as the feature partition, and then re-obtain the clustering coefficient corresponding to this feature partition, and so on, until the optimization of the feature partition is completed.

[0121] The above are only the preferred embodiments of the present invention, and do not impose any form of limitation on the present invention. Although the present invention has been disclosed above in the preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some changes or modifications to equivalent embodiments of equivalent changes within the scope of the technical solution of the present invention by using the above-disclosed technical content. However, as long as it does not depart from the content of the technical solution of the present invention, any modification or equivalent replacement made to the above embodiments based on the technical essence of the present invention still falls within the scope of the technical solution of the present invention.

Claims

1. A sonar imaging optimization method for underwater environmental condition monitoring, characterized in that, Including the following steps: Step S1: Arrange the sonar front ends at different positions within the target ocean area, and construct a corresponding sonar front end matrix according to the arranged sonar front ends; Step S2: Transmit corresponding acoustic wave signals to the target ocean area through the sonar front end matrix, and receive the corresponding reflected signals; Step S3: Process the reflected signals received by the sonar front end matrix to complete the construction of the underwater simulation image of the target ocean area; Step S4: Perform sub-region optimization on the constructed underwater simulation image, and generate corresponding underwater factor labels within the underwater simulation image according to the optimization results.

2. The sonar imaging optimization method for underwater environmental condition monitoring according to claim 1, wherein The process of arranging the sonar front ends at different positions within the target ocean area and constructing a corresponding sonar front end matrix includes: Construct a three-dimensional coordinate system, select the target ocean area, generate an ocean plane model according to the scope of the target ocean area, and map the ocean plane model into the three-dimensional coordinate system; Set several sonar front ends within the target ocean area; Select the location of any sonar front end as the origin and map it into the three-dimensional coordinate system, and map the locations of the other set sonar front ends into the three-dimensional coordinate system; Obtain the acoustic wave signal emission angle range corresponding to each sonar front end, and generate an acoustic wave signal coverage area within the three-dimensional coordinate system according to the acoustic wave signal emission angle range; According to the number of sonar front ends, construct a sonar front end matrix composed of several blank matrix units, and associate each sonar front end with a blank matrix unit respectively; Import the location coordinates of the sonar front ends into the corresponding blank matrix units.

3. The sonar imaging optimization method for underwater environmental condition monitoring according to claim 2, wherein, The process of transmitting corresponding acoustic wave signals to the target ocean area through the sonar front end matrix and receiving the corresponding reflected signals includes: The sonar front ends emit acoustic wave signals of different frequencies according to the corresponding acoustic wave signal emission angle ranges, and each sonar front end emits acoustic wave signals in a rotation mode. After all sonar front ends complete the rotation, it is recorded as a rotation period, and the reflected signals corresponding to each acoustic wave signal within the rotation period are collected; Record the signal emission time and signal parameters of the acoustic wave signals emitted by each acoustic wave signal; Summarize the acoustic wave signals and reflected signals within a rotation period to obtain a corresponding signal data set.

4. A sonar imaging optimization method for underwater environmental condition monitoring according to claim 3, characterized in that, The process of processing the reflected signals received by the sonar front end matrix includes: Construct a two-dimensional coordinate system of time versus signal parameters corresponding to each sonar front end; Set a corresponding frequency fluctuation range according to the frequency table; Filter the reflected signals received by each sonar front end through the set frequency fluctuation range; Map the filtered reflected signals into the corresponding coordinate system to obtain a spectrogram corresponding to each sonar terminal; Input the obtained spectrograms into the trained convolutional neural network model, identify the spectral features within each spectrogram through the convolutional neural network model, and convert the identified spectral features into corresponding underwater factor features.

5. The sonar imaging optimization method for underwater environmental condition monitoring according to claim 4, wherein The training process of the convolutional neural network model is: Collect sample data, where the sample data is the spectral signal corresponding to the reflected signal received after the sonar terminal emits acoustic wave signals according to the frequencies in the frequency table. Associate the corresponding true features with each sample data, and input the collected sample data into the convolutional neural network model; The convolutional neural network model outputs the corresponding predicted features according to the input sample data; Input the output predicted features and true features into the loss function to obtain the feature loss value; Repeat the above operations until the set specific number of training times or the feature loss value obtained by the loss function is reached, thus completing the training process of the convolutional neural network model.

6. The sonar imaging optimization method for underwater environmental condition monitoring according to claim 5, characterized in that The process of constructing the underwater simulation image of the target ocean area includes: Select the sonar front end as the reference end in sequence, and according to the signal frequency of the acoustic wave signal emitted by the reference end, obtain the spectrogram of the reflection signal of each sonar front end corresponding to the signal frequency; Record the time corresponding to the recognized spectral feature as t j ; Record the acoustic wave signal emission time of the sonar terminal as t0; Then obtain the distance from the sonar terminal to the underwater factor feature corresponding to the spectral feature; Select the moment corresponding to the spectral feature of any other sonar terminal, and further obtain the azimuth of the underwater factor feature corresponding to the spectral feature; Generate the corresponding underwater feature model in the three-dimensional coordinate system according to the obtained azimuth and distance of the underwater factor feature; Complete the positioning of all spectral features corresponding to the underwater factor features in the spectrogram, and generate the corresponding underwater feature model.

7. The sonar imaging optimization method for underwater environmental condition monitoring according to claim 6, wherein The process of performing regional optimization on the constructed underwater simulation image and generating the corresponding underwater factor labels in the underwater simulation image according to the optimization results includes: Mark the spatial coordinates of the obtained underwater feature models corresponding to each sonar front end in the three-dimensional coordinate system; Obtain the two spatial coordinates with the largest marked straight-line distance, and generate the corresponding connection line between the two spatial coordinates; Generate the corresponding spherical space with the connection line between the two connected spatial coordinates as the diameter, and mark the generated spherical space as the corresponding feature partition; Obtain the center point coordinates of the spherical space, and obtain the straight-line distance from the center point to each spatial coordinate in the spherical space; Furthermore, obtain the clustering coefficient Jx of this feature partition; If Jx≥0, it means that the clustering degree of this feature partition is high, then take the center point of this feature partition as the clustering point of this underwater feature model, and cluster the spatial coordinates of all underwater feature models in the feature partition to the clustering point, thus completing the optimization of this feature partition and generating the underwater factor label corresponding to this clustering point; When Jx<0, it means that the clustering degree of this feature partition is low, then eliminate the two connected spatial coordinates, and select the two spatial coordinates with the largest straight-line distance from the remaining spatial coordinates to generate a new spherical space as the feature partition, and then re-obtain the clustering coefficient corresponding to this feature partition, and so on until the optimization of the feature partition is completed.

Images