A multi-dimensional underwater biodiversity automated monitoring method, system and terminal

Through the multi-dimensional underwater biodiversity automated monitoring method, the correlation network model is trained using environmental DNA, underwater image and sound data, and the problems of low sampling frequency and insufficient accuracy in marine biodiversity monitoring are solved, achieving high accuracy and long-term Dynamic monitoring and automated early warning.

CN118378216BActive Publication Date: 2024-11-12GUANGDONG LABORATORY OF SOUTHERN OCEAN SCIENCE AND ENGINEERING (GUANGZHOU) +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410815925.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-06-24
Publication Date
2024-11-12
Estimated Expiration
2044-06-24

AI Technical Summary

Technical Problem

The existing technology has low sampling frequency in marine biodiversity monitoring, resulting in low monitoring accuracy and susceptible to hardware, lack of accuracy verification, resulting in low accuracy of monitoring results.

Method used

Using multi-dimensional underwater biodiversity automation monitoring method, we can pre-process environmental DNA data, underwater image data and underwater sound data, train a correlation network model, combine multiple data to perform biodiversity analysis, and achieve high spatial and temporal resolution rate monitoring.

Benefits of technology

It has achieved high-precision and long-term dynamic monitoring of marine biodiversity, timely discover ecological environment changes, predict ecosystem development, and provide automated early warning and protection.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118378216B_ABST
    Figure CN118378216B_ABST
Patent Text Reader

Abstract

The present invention discloses a multi-dimensional underwater biodiversity automatic monitoring method, system and terminal, the method comprising: obtaining environmental DNA data, underwater image data and underwater sound data of a target sea area, and performing preprocessing to obtain target environmental DNA data, target underwater image data and target underwater sound data; performing model training according to the target environmental DNA data, target underwater image data and target underwater sound data to obtain a target correlation analysis model; obtaining the underwater image data and underwater sound data to be tested of the sea area to be tested, inputting them into the target correlation analysis model, and outputting underwater biodiversity monitoring results. The present invention can perform long-term dynamic monitoring of the complex underwater ecological environment, and timely obtain effective information of marine organisms, so as to timely discover potential signals before changes in the ecological environment, predict the development of the ecosystem, and thus realize automatic early warning and protection of marine biodiversity.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of underwater biological monitoring, and in particular to a multi-dimensional underwater biodiversity automatic monitoring method, system, terminal and storage medium. Background Art

[0002] With global climate change and other disturbances, marine biodiversity and its sustainable use are under serious threat. The protection of marine biodiversity has gradually become an important concern. Scientific and reasonable diversity protection measures are formulated using a variety of means to monitor regional oceans. At present, only a limited number of biological and ecological indicators can be monitored in short time series and at low resolution. This monitoring method is not conducive to understanding the dynamics of marine ecosystems. However, due to the large number of marine organisms, the complexity of monitoring content, and the constraints of factors such as weather and sea conditions, traditional survey methods (such as trawling) require a lot of manpower and material resources, and are also subject to the constraints of factors such as weather and sea conditions, thereby limiting the acquisition of marine biodiversity data and failing to achieve the role of continuous and efficient monitoring. In addition, the survey results can only reflect the biodiversity at a certain point in time, so there is a lack of continuity and effectiveness of biological data.

[0003] With the innovation and progress of technology, environmental DNA biodiversity detection, underwater optical imaging technology and underwater acoustic technology have begun to be applied to marine biodiversity monitoring. Compared with traditional survey methods, they have some advantages in cost or monitoring time. However, whether it is environmental DNA biodiversity detection, underwater optical imaging technology or underwater acoustic technology, they all have certain limitations, and it is difficult to achieve high-precision and high-frequency monitoring of biodiversity with a single method; for example, although environmental DNA biodiversity detection has high detection accuracy, the sampling frequency is relatively limited, which makes DNA extraction and detection accuracy easily affected; underwater biological image biodiversity monitoring is suitable for monitoring medium and large animals within a short distance. According to statistical data over a certain period of time, the number and distribution of some animal populations are evaluated. However, because underwater cameras are easily affected by factors such as image resolution, sea conditions and instrument hardware, the monitoring of biodiversity is obviously more one-sided and lacks accuracy verification, resulting in low accuracy of monitoring results.

[0004] Therefore, the prior art still needs to be improved and developed. Summary of the invention

[0005] The main purpose of the present invention is to provide a multi-dimensional underwater biodiversity automated monitoring method, system and terminal, aiming to solve the problem that the existing technology has a low sampling frequency for marine biodiversity monitoring, resulting in low monitoring accuracy, and is easily affected by related hardware, resulting in one-sided monitoring content, and lacks accuracy verification, resulting in low accuracy of monitoring results.

[0006] To achieve the above object, the present invention provides a multi-dimensional underwater biodiversity automated monitoring method, the multi-dimensional underwater biodiversity automated monitoring method comprising the following steps:

[0007] Acquire environmental DNA data, underwater image data and underwater sound data of a target sea area, and pre-process the environmental DNA data, underwater image data and underwater sound data to obtain target environmental DNA data, target underwater image data and target underwater sound data;

[0008] The created correlation network training model is trained according to the target environment DNA data, the target underwater image data and the target underwater sound data to obtain a target correlation analysis model;

[0009] The underwater image data and underwater sound data of the sea area to be tested are obtained, the underwater image data and underwater sound data to be tested are input into the target correlation analysis model, and the underwater biodiversity monitoring result is output.

[0010] Optionally, the multi-dimensional underwater biodiversity automated monitoring method, wherein the step of acquiring environmental DNA data, underwater image data and underwater sound data of a target sea area, and pre-processing the environmental DNA data, the underwater image data and the underwater sound data to obtain target environmental DNA data, target underwater image data and target underwater sound data, specifically comprises:

[0011] Setting a sampling depth, performing multi-channel sampling on the biological DNA data of the target sea area according to the sampling depth to obtain environmental DNA data, performing image acquisition on the underwater organisms in the target sea area according to the sampling depth to obtain underwater image data, and performing sound acquisition on the target sea area according to the sampling depth to obtain underwater sound data;

[0012] The environmental DNA data, the underwater image data and the underwater sound data are subjected to data review to obtain corresponding abnormal values, and the abnormal values ​​corresponding to the environmental DNA data, the underwater image data and the underwater sound data are corrected to obtain target environmental DNA data, target underwater image data and target underwater sound data, respectively. The abnormal values ​​include missing values, error values ​​and abnormal outliers.

[0013] Optionally, the multi-dimensional underwater biodiversity automated monitoring method, wherein the created correlation network training model is trained according to the target environmental DNA data, the target underwater image data and the target underwater sound data to obtain a target correlation analysis model, specifically comprising:

[0014] Creating a relevant network training model, performing species diversity analysis on the target environmental DNA data to obtain a biodiversity result, using the biodiversity result as a validation set, and using the target underwater image data and the target underwater sound data as a training set, wherein the training set includes a plurality of training groups;

[0015] Inputting a training group in the training set into the relevant network training model, performing species recognition processing on the target underwater image data to obtain a biological recognition result, and performing sound recognition analysis on the target underwater sound data to obtain a sound analysis result;

[0016] Analyze the biometric recognition result and the sound analysis result to obtain a predicted biodiversity result, and calibrate model parameters of the relevant network training model according to the validation set and the predicted biodiversity result;

[0017] The next training group is input into the related network training model until the training status of the related network training model meets the preset conditions, thereby obtaining a trained target correlation analysis model.

[0018] Optionally, the multi-dimensional underwater biodiversity automated monitoring method, wherein the performing of species diversity analysis on the target environmental DNA data to obtain biodiversity results, specifically includes:

[0019] Acquire a target target region of the target environmental DNA data, and perform primer amplification processing on the target target region to obtain an amplicon sequence;

[0020] The amplicon sequence is subjected to high-throughput sequencing to obtain a sequencing result, and data analysis is performed on the sequencing result to obtain a biodiversity result, wherein the biodiversity result includes species composition and biodiversity information.

[0021] Optionally, the multi-dimensional underwater biodiversity automated monitoring method, wherein the step of performing species identification processing on the target underwater image data to obtain a biological identification result, specifically includes:

[0022] Performing image preprocessing on the target underwater image data to obtain a target image, wherein the image preprocessing includes color correction processing, defogging processing and filtering processing;

[0023] Based on the trained image recognition model, the target image is subjected to feature extraction to obtain multiple feature information, all the feature information is fused to obtain the target feature, the organisms in the target image are marked according to the target feature to obtain the number of organisms, and the number of organisms is classified into species to obtain the biological recognition result.

[0024] Optionally, the multi-dimensional underwater biodiversity automated monitoring method, wherein the step of performing sound recognition analysis on the target underwater sound data to obtain a sound analysis result, specifically includes:

[0025] Performing high-frequency signal detection on the target underwater sound data according to a first preset signal detection threshold to obtain a target high-frequency signal, performing feature extraction on the target high-frequency signal to obtain a target high-frequency feature, and obtaining acoustic complexity and sound diversity according to the target high-frequency feature;

[0026] Performing low-frequency signal detection on the target underwater sound data according to a second preset signal detection threshold to obtain a target low-frequency signal, performing feature extraction on the target low-frequency signal to obtain a target low-frequency feature, and obtaining sound richness and species type according to the target low-frequency feature;

[0027] Calculating the sound pressure of the target underwater sound data to obtain the sound pressure intensity and the octave band sound pressure level;

[0028] A sound analysis result is obtained according to the acoustic complexity, the sound diversity, the sound richness, the species type, the sound pressure intensity and the octave band sound pressure level.

[0029] Optionally, the multi-dimensional underwater biodiversity automated monitoring method, wherein the analyzing the biometric identification result and the sound analysis result to obtain the predicted biodiversity result, specifically includes:

[0030] Obtaining the real biodiversity results of the target sea area, performing correlation analysis on the biometric identification results and the sound analysis results with the biodiversity results to obtain a first analysis result, and performing correlation analysis on the biometric identification results and the sound analysis results with the real biodiversity results to obtain a second analysis result;

[0031] Data screening is performed on the first analysis result and the second analysis result to obtain a biodiversity assessment factor, and biodiversity prediction is performed based on the biodiversity assessment factor to obtain a predicted biodiversity result.

[0032] Optionally, in the multi-dimensional underwater biodiversity automated monitoring method, the multi-dimensional underwater biodiversity automated monitoring system comprises:

[0033] A data acquisition module is used to acquire environmental DNA data, underwater image data and underwater sound data of a target sea area, and pre-process the environmental DNA data, underwater image data and underwater sound data to obtain target environmental DNA data, target underwater image data and target underwater sound data;

[0034] A data processing module, used for training the created correlation network training model according to the target environment DNA data, the target underwater image data and the target underwater sound data to obtain a target correlation analysis model;

[0035] The biological monitoring module is used to obtain underwater image data and underwater sound data to be tested, input the underwater image data and underwater sound data to be tested into the target correlation analysis model, and output the target biodiversity results to complete the monitoring of underwater biodiversity.

[0036] In addition, to achieve the above-mentioned purpose, the present invention also provides a terminal, wherein the terminal includes: a memory, a processor, and a program stored in the memory and executable on the processor, and when the program is executed by the processor, the steps of the multi-dimensional underwater biodiversity automated monitoring method as described above are implemented.

[0037] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a multi-dimensional underwater biodiversity automated monitoring program, and when the multi-dimensional underwater biodiversity automated monitoring program is executed by a processor, the steps of the multi-dimensional underwater biodiversity automated monitoring method as described above are implemented.

[0038] In the present invention, environmental DNA data, underwater image data and underwater sound data of the target sea area are obtained, and the environmental DNA data, underwater image data and underwater sound data are preprocessed to obtain target environmental DNA data, target underwater image data and target underwater sound data; the created correlation network training model is trained according to the target environmental DNA data, the target underwater image data and the target underwater sound data to obtain a target correlation analysis model; the underwater image data and underwater sound data to be tested of the sea area to be tested are obtained, and the underwater image data and underwater sound data to be tested are input into the target correlation analysis model, and the underwater biodiversity monitoring result is output. When the present invention is used for biodiversity monitoring, it is not necessary to capture marine life, so it will not cause damage to the monitored ecosystem. It is also suitable for long-term dynamic monitoring of complex marine ecological environment, and timely automatic identification of effective information of marine life, so as to timely discover potential signals before changes in the marine ecological environment, to predict the development of the corresponding marine ecosystem, and realize automatic early warning and protection of marine biodiversity. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a flow chart of a preferred embodiment of the multi-dimensional underwater biodiversity automated monitoring method of the present invention;

[0040] Figure 2 It is a schematic diagram of a device for realizing a multi-dimensional underwater biodiversity automated monitoring method in the present invention;

[0041] Figure 3 It is a flow chart of the processing process of the high-definition underwater optical intelligent recognition subsystem in a preferred embodiment of the present invention;

[0042] Figure 4 is a flow chart of the processing process of the intelligent marine biological sound diversity monitoring subsystem in a preferred embodiment of the present invention;

[0043] Figure 5 It is a schematic diagram of the principle of a preferred embodiment of the multi-dimensional underwater biodiversity automatic monitoring system of the present invention;

[0044] Figure 6 It is a schematic diagram of the operating environment of a preferred embodiment of the terminal of the present invention. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical solution and advantages of the present invention clearer and more specific, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0046] It should be noted that if the embodiments of the present invention involve directional indications (such as up, down, left, right, front, back...), the directional indications are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0047] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present invention, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or suggesting their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0048] The multi-dimensional underwater biodiversity automated monitoring method described in the preferred embodiment of the present invention is as follows: Figure 1 As shown, the multi-dimensional underwater biodiversity automated monitoring method comprises the following steps:

[0049] Step S10, obtaining environmental DNA data, underwater image data and underwater sound data of the target sea area, preprocessing the environmental DNA data, the underwater image data and the underwater sound data to obtain target environmental DNA data, target underwater image data and target underwater sound data.

[0050] The step S10 comprises:

[0051] Step S11, setting a sampling depth, performing multi-channel sampling on the biological DNA data of the target sea area according to the sampling depth to obtain environmental DNA data, performing image acquisition on the underwater organisms in the target sea area according to the sampling depth to obtain underwater image data, and performing sound acquisition on the target sea area according to the sampling depth to obtain underwater sound data;

[0052] Step S12, perform data review on the environmental DNA data, the underwater image data and the underwater sound data to obtain corresponding abnormal values, and correct the abnormal values ​​corresponding to the environmental DNA data, the underwater image data and the underwater sound data to obtain target environmental DNA data, target underwater image data and target underwater sound data, wherein the abnormal values ​​include missing values, error values ​​and abnormal outliers.

[0053] Specifically, in the embodiment of the present invention, Figure 2As shown, a multi-dimensional underwater biodiversity automated monitoring method is realized by means of the device deployed in the target sea area, which mainly includes an environmental DNA biodiversity monitoring subsystem, a high-definition underwater optical intelligent recognition subsystem, an intelligent marine biological sound diversity monitoring subsystem and a cross-domain power supply and data transmission subsystem; specifically, the environmental DNA biodiversity monitoring subsystem includes a plurality of sampling channels, each of which contains a protected filter membrane, through which the eDNA (DNA left over from the interaction between organisms and the environment, i.e., environmental DNA) shed by marine organisms can be continuously collected, and the sampling depth is set, for example, 1 meter underwater, and the environmental DNA biodiversity monitoring subsystem performs multi-channel automatic continuous sampling of the biological DNA data of the target sea area at the sampling depth, and obtains the environmental DNA data through the filter membrane of each channel without filtering or other processing, wherein the environmental DNA data is obtained by extracting all the filter membranes using the Qiagen DNeasy tissue and blood DNA extraction kit, and then performing PCR (Polymerase Chain Reaction) Reaction, polymerase chain reaction) amplification and high-throughput sequencing; the high-definition underwater optical intelligent recognition subsystem is a target detection and efficient recognition system based on a deep convolutional neural network model. It uses a high-definition underwater camera to collect images of underwater organisms in the target sea area to obtain underwater image data. Subsequently, the underwater organisms in the underwater image data are detected and identified through technical means such as data denoising, feature extraction and fusion, and multimodal analysis; the intelligent marine biological sound diversity monitoring subsystem includes a sound collection module, a signal processing module, and a signal transmission module. The sound collection module can realize 50 Hz to 150 kHz bandwidth, the signal processing module can realize intelligent processing and analysis of the collected underwater sound data; the signal transmission module can transmit the analysis results in real time to realize real-time monitoring of relevant information; the cross-domain power supply and data transmission subsystem includes a wave energy power generation buoy, a surface observation unit and an underwater observation unit, which is a new type of ocean observation-communication-energy integrated system integrating wave energy power generation, environmental element observation, broadband information transmission and comprehensive display and control. The wave energy power generation buoy uses the energy of waves as the power to produce electricity, and provides electricity for the environmental DNA biodiversity monitoring subsystem, the high-definition underwater optical intelligent recognition subsystem and the intelligent marine biological sound diversity monitoring subsystem. The surface observation unit includes a wave sensor, a multi-parameter water quality meter, a data acquisition module and a communication module, etc., which are used to complete the functions of node information aggregation, storage and transmission. The underwater observation unit mainly includes a dynamic cable system and a seabed base, and the seabed base is used to provide a carrying platform; the cross-domain power supply and data transmission subsystem can also complete the collection of basic environmental parameters such as chlorophyll, dissolved oxygen, temperature, salinity and depth.

[0054] After obtaining the environmental DNA data, the underwater image data and the underwater sound data, it is necessary to conduct a data review on the environmental DNA data, the underwater image data and the underwater sound data to promptly discover corresponding abnormal values, which include missing values, error values ​​and abnormal outliers. The abnormal values ​​corresponding to the environmental DNA data, the underwater image data and the underwater sound data are corrected to obtain target environmental DNA data, target underwater image data and target underwater sound data, so that the values ​​of the acquired data are normalized and standardized, thereby ensuring the comparability of the data.

[0055] Step S20: training the created correlation network training model according to the target environment DNA data, the target underwater image data and the target underwater sound data to obtain a target correlation analysis model.

[0056] Specifically, in an embodiment of the present invention, after obtaining the target environmental DNA data, the target underwater image data and the target underwater sound data, first, the target environmental DNA data is PCR amplified and high-throughput sequenced by the environmental DNA biodiversity monitoring subsystem, specifically, the 12S rRNA region, the 18S rRNA region and the COI region (cytochrome oxidase I region) are used as target regions, and the primers of the target target regions are PCR amplified to obtain amplicon sequences, and the amplicon sequences are subjected to high-throughput sequencing to obtain sequencing results, and the sequencing results are subjected to data analysis to obtain biodiversity results, wherein the biodiversity results include species composition and biodiversity information; although the environmental DNA biodiversity monitoring subsystem has high detection accuracy for environmental DNA biodiversity detection, there are problems with sample collection and storage, and the sampling frequency is relatively limited, so in the present invention, the environmental DNA biodiversity monitoring subsystem is used as a calibration method; and both the high-definition underwater optical intelligent recognition subsystem and the intelligent marine biological sound diversity monitoring subsystem have Some problems, for example, the accuracy of biodiversity monitoring is low. Therefore, in the present invention, the diversity information collected by the environmental DNA biodiversity monitoring subsystem, the high-definition underwater optical intelligent recognition subsystem and the intelligent marine biological sound diversity monitoring subsystem is subjected to multi-dimensional information fusion analysis, that is, the underwater biodiversity is monitored in real time through the target correlation analysis model. The target correlation analysis model is obtained by training the relevant network training model created by combining the high-definition underwater optical intelligent recognition subsystem and the intelligent marine biological sound diversity monitoring subsystem. The biodiversity results obtained by the environmental DNA biodiversity monitoring subsystem are used as a verification set, and the target underwater image data and the target underwater sound data are used as a training set, and the training set includes multiple training groups.

[0057] A training group in the training set is input into the relevant network training model, and the high-definition underwater optical intelligent recognition subsystem of the relevant network training model performs species recognition processing on the target underwater image data to obtain a biological recognition result. The specific processing process is as follows: Figure 3 As shown, after the target underwater image data is acquired by a high-definition underwater camera, the target underwater image data is first subjected to data denoising. In the data denoising process, for the target underwater image data acquired in the underwater low-light and high-noise complex environment, the target underwater image data is subjected to color correction, denoising, defogging and filtering removal respectively by a color correction processing module, a defogging processing module and a filtering processing module to obtain a target image, so as to improve the image data quality and suppress redundant information, thereby improving the recognition accuracy and realizing the subsequent reliable labeling of the target organism; subsequently, the created image recognition training model is subjected to image training by an underwater image database module and a deep learning module to obtain an image recognition model, wherein the image recognition model includes a target detection module and an object classification module to perform target detection and species classification on the image, and can realize real-time biological information in the image at a speed of 0.33 frames per second. Detection and classification improve the efficiency and accuracy of biometric identification; then, the target image is input into a trained image recognition model, and the image recognition model extracts features from the target image to obtain multiple feature information, such as color, shape, contour and texture, so as to mine the inter-class information of the target in the same perspective and the inter-class information between different perspectives, and then for the subsequent reconstruction of high-level features (i.e., target features), all feature information is fused to obtain target features, and then the target image is annotated by the target features to perform tasks such as biological individual tracking, biological individual counting and morphological measurement, and classification of multiple biological individuals, biological behaviors and species; specifically, the organisms in the target image are marked according to the target features by the target detection module to obtain the number of organisms, and the number of organisms is classified by species by the object classification module to obtain the biometric identification result.

[0058] The intelligent marine biological sound diversity monitoring subsystem of the relevant network training model performs sound recognition analysis on the target underwater sound data to obtain a sound analysis result. The specific processing process is as follows: Figure 4As shown, after collecting the target underwater sound data, it is necessary to perform signal analysis on the target underwater sound data. Since the source of the ocean sound signal is complex, the signal parameters are relatively different, and the signal analysis content requirements are relatively large, the traditional data processing process requires too high hardware performance of the equipment, and the data processing time is too long, which is not conducive to rapid data processing; in the embodiment of the present invention, in view of the large difference between marine organisms and environmental sound signals and the diverse task requirements, the corresponding sound signal processing work is assigned to three different processing modules in the signal processing module to achieve rapid data processing and rapid extraction of parameter features, wherein the three different processing modules are a marine organism high-frequency sound signal processing module, a marine organism low-frequency sound signal processing module and a marine environmental noise processing module; then, the target underwater sound data is simultaneously input into these three modules for analysis; for the marine organism high-frequency sound signal processing module, in order to improve the signal-to-noise ratio and reduce the interference of the reflected signal, a higher filter can be set, for example, 30 kHz or above, and set the corresponding signal detection threshold, that is, the first preset signal detection threshold, to avoid noise interference, which not only facilitates the rapid detection of high-frequency signals and parameter feature extraction, but also reduces the workload of the system, and performs high-frequency signal detection on the target underwater sound data according to the first preset signal detection threshold to obtain the target high-frequency signal, and performs feature extraction on the target high-frequency signal to obtain the target high-frequency feature, and obtains the acoustic complexity and sound diversity according to the target high-frequency feature; for the marine biological low-frequency sound signal processing module, the interference of high-frequency noise signals and low-frequency ocean natural background noise signals is eliminated by bandpass filtering, and the signal detection threshold is adjusted according to the intensity of the target underwater sound data to obtain the second preset signal detection threshold, which not only eliminates the influence of noise to the greatest extent. , and also improves the signal detection rate and the accuracy of feature extraction; according to the second preset signal detection threshold, low-frequency signal detection is performed on the target underwater sound data to obtain the target low-frequency signal, feature extraction is performed on the target low-frequency signal to obtain the target low-frequency feature, and the sound richness and species type are obtained according to the target low-frequency feature; for the marine environmental noise signal processing module, in order to ensure the true description of the information characteristics of the marine environmental noise, there is no need to perform filtering processing, and the sound pressure of the target underwater sound data is directly calculated to obtain the sound pressure intensity and the octave sound pressure level, for example, the 1 / 3 octave sound pressure level; then the sound analysis result is obtained according to the acoustic complexity, the sound diversity, the sound richness, the species type, the sound pressure intensity and the octave sound pressure level.

[0059] Afterwards, the biometric identification results and the sound analysis results are analyzed to obtain the real biodiversity results of the target sea area (i.e., the biodiversity results of artificial surveys); according to the accuracy of the biodiversity survey, the biometric identification results, the sound analysis results, the biodiversity results and the real biodiversity results are subjected to correlation analysis; the biodiversity results corresponding to the environmental DNA data are used as the primary correlation analysis, and the real biodiversity results are used as the secondary correlation analysis, that is, the biometric identification results and the sound analysis results are subjected to correlation analysis with the biodiversity results to obtain a first analysis result, and the biometric identification results and the sound analysis results are subjected to correlation analysis with the real biodiversity results to obtain a second analysis result; data screening is performed on the first analysis result and the second analysis result to obtain a parameter group with the highest correlation with the actual results in the indicators of the optical data and the acoustic data, and this parameter group is used as a biodiversity assessment factor, thereby realizing correlation analysis of multi-angle collaboration and multimodal fusion; and biodiversity prediction is then performed according to the biodiversity assessment factor to obtain a predicted biodiversity result.

[0060] The model parameters of the relevant network training model are corrected according to the validation set and the predicted biodiversity results to ensure the generalization ability of the subsequently obtained target correlation analysis model in real scenarios; then the next training group is input into the relevant network training model until the training status of the relevant network training model meets the preset conditions, thereby obtaining a trained target correlation analysis model, wherein the preset requirements can be determined according to the accuracy of the target correlation analysis model, which is not described in detail here.

[0061] Step S30, obtaining underwater image data and underwater sound data of the sea area to be tested, inputting the underwater image data and underwater sound data to be tested into the target correlation analysis model, and outputting underwater biodiversity monitoring results.

[0062] Specifically, underwater image data and underwater sound data to be tested of the sea area to be tested are obtained, the underwater image data and underwater sound data to be tested are input into the target correlation analysis model, and the underwater biodiversity monitoring results are output to monitor the underwater biodiversity in real time. The present invention can monitor marine biodiversity with high temporal and spatial resolution and without loss, does not need to capture marine life, and will not cause damage to the ecosystem of the monitored marine life. It also uses wave energy for autonomous power supply and does not need to be connected to the shore by cable. It can realize long-term monitoring of biodiversity underwater, and is suitable for long-term dynamic monitoring of complex marine ecological environment, and can automatically identify effective information of marine life in time, so as to timely discover potential signals before changes in the marine ecological environment, to predict the development of the corresponding marine ecosystem, and realize automatic early warning and protection of marine biodiversity.

[0063] Furthermore, if Figure 5 As shown, based on the above multi-dimensional underwater biodiversity automated monitoring method, the present invention also provides a multi-dimensional underwater biodiversity automated monitoring system, the multi-dimensional underwater biodiversity automated monitoring system comprising:

[0064] The data acquisition module 51 is used to acquire the environmental DNA data, underwater image data and underwater sound data of the target sea area, and pre-process the environmental DNA data, the underwater image data and the underwater sound data to obtain the target environmental DNA data, the target underwater image data and the target underwater sound data;

[0065] A data processing module 52 is used to train the created correlation network training model according to the target environment DNA data, the target underwater image data and the target underwater sound data to obtain a target correlation analysis model;

[0066] The biological monitoring module 53 is used to obtain underwater image data and underwater sound data of the sea area to be tested, input the underwater image data and underwater sound data to be tested into the target correlation analysis model, and output underwater biodiversity monitoring results.

[0067] Furthermore, if Figure 6 As shown, based on the above multi-dimensional underwater biodiversity automatic monitoring method, the present invention also provides a terminal accordingly, and the terminal includes a processor 10, a memory 20 and a display 30. Figure 6 Only some components of the terminal are shown, but it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0068] In some embodiments, the memory 20 may be an internal storage unit of the terminal, such as a hard disk or memory of the terminal. In other embodiments, the memory 20 may also be an external storage device of the terminal, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the terminal. Furthermore, the memory 20 may also include both an internal storage unit of the terminal and an external storage device. The memory 20 is used to store application software and various types of data installed on the terminal, such as the program code of the installation terminal. The memory 20 may also be used to temporarily store data that has been output or is to be output. In one embodiment, a multi-dimensional underwater biodiversity automated monitoring program 40 is stored on the memory 20, and the multi-dimensional underwater biodiversity automated monitoring program 40 can be executed by the processor 10, thereby realizing the multi-dimensional underwater biodiversity automated monitoring method in the present application.

[0069] In some embodiments, the processor 10 may be a central processing unit (CPU), a microprocessor or other data processing chip, used to run the program code or process data stored in the memory 20, such as executing the multi-dimensional underwater biodiversity automated monitoring method.

[0070] In some embodiments, the display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, etc. The display 30 is used to display information on the terminal and to display a visual user interface. The components 10-30 of the terminal communicate with each other via a system bus.

[0071] In one embodiment, when the processor 10 executes the program 40 for multi-dimensional underwater biodiversity automated monitoring in the memory 20, the following steps are implemented:

[0072] Acquire environmental DNA data, underwater image data and underwater sound data of a target sea area, and pre-process the environmental DNA data, underwater image data and underwater sound data to obtain target environmental DNA data, target underwater image data and target underwater sound data;

[0073] The created correlation network training model is trained according to the target environment DNA data, the target underwater image data and the target underwater sound data to obtain a target correlation analysis model;

[0074] The underwater image data and underwater sound data of the sea area to be tested are obtained, the underwater image data and underwater sound data to be tested are input into the target correlation analysis model, and the underwater biodiversity monitoring result is output.

[0075] The step of acquiring the environmental DNA data, underwater image data and underwater sound data of the target sea area, and preprocessing the environmental DNA data, underwater image data and underwater sound data to obtain target environmental DNA data, target underwater image data and target underwater sound data specifically includes:

[0076] Setting a sampling depth, performing multi-channel sampling on the biological DNA data of the target sea area according to the sampling depth to obtain environmental DNA data, performing image acquisition on the underwater organisms in the target sea area according to the sampling depth to obtain underwater image data, and performing sound acquisition on the target sea area according to the sampling depth to obtain underwater sound data;

[0077] The environmental DNA data, the underwater image data and the underwater sound data are subjected to data review to obtain corresponding abnormal values, and the abnormal values ​​corresponding to the environmental DNA data, the underwater image data and the underwater sound data are corrected to obtain target environmental DNA data, target underwater image data and target underwater sound data, respectively. The abnormal values ​​include missing values, error values ​​and abnormal outliers.

[0078] The step of training the created correlation network training model according to the target environment DNA data, the target underwater image data and the target underwater sound data to obtain the target correlation analysis model specifically includes:

[0079] Creating a relevant network training model, performing species diversity analysis on the target environmental DNA data to obtain a biodiversity result, using the biodiversity result as a validation set, and using the target underwater image data and the target underwater sound data as a training set, wherein the training set includes a plurality of training groups;

[0080] Inputting a training group in the training set into the relevant network training model, performing species recognition processing on the target underwater image data to obtain a biological recognition result, and performing sound recognition analysis on the target underwater sound data to obtain a sound analysis result;

[0081] Analyze the biometric recognition result and the sound analysis result to obtain a predicted biodiversity result, and calibrate model parameters of the relevant network training model according to the validation set and the predicted biodiversity result;

[0082] The next training group is input into the related network training model until the training status of the related network training model meets the preset conditions, thereby obtaining a trained target correlation analysis model.

[0083] The performing of species diversity analysis on the target environmental DNA data to obtain biodiversity results specifically includes:

[0084] Acquire a target target region of the target environmental DNA data, and perform primer amplification processing on the target target region to obtain an amplicon sequence;

[0085] The amplicon sequence is subjected to high-throughput sequencing to obtain a sequencing result, and data analysis is performed on the sequencing result to obtain a biodiversity result, wherein the biodiversity result includes species composition and biodiversity information.

[0086] The performing of species identification processing on the target underwater image data to obtain a biological identification result specifically includes:

[0087] Performing image preprocessing on the target underwater image data to obtain a target image, wherein the image preprocessing includes color correction processing, defogging processing and filtering processing;

[0088] Based on the trained image recognition model, the target image is subjected to feature extraction to obtain multiple feature information, all the feature information is fused to obtain the target feature, the organisms in the target image are marked according to the target feature to obtain the number of organisms, and the number of organisms is classified into species to obtain the biological recognition result.

[0089] The performing of sound recognition analysis on the target underwater sound data to obtain a sound analysis result specifically includes:

[0090] Performing high-frequency signal detection on the target underwater sound data according to a first preset signal detection threshold to obtain a target high-frequency signal, performing feature extraction on the target high-frequency signal to obtain a target high-frequency feature, and obtaining acoustic complexity and sound diversity according to the target high-frequency feature;

[0091] Performing low-frequency signal detection on the target underwater sound data according to a second preset signal detection threshold to obtain a target low-frequency signal, performing feature extraction on the target low-frequency signal to obtain a target low-frequency feature, and obtaining sound richness and species type according to the target low-frequency feature;

[0092] Calculating the sound pressure of the target underwater sound data to obtain the sound pressure intensity and the octave band sound pressure level;

[0093] A sound analysis result is obtained according to the acoustic complexity, the sound diversity, the sound richness, the species type, the sound pressure intensity and the octave band sound pressure level.

[0094] The step of analyzing the biometric identification result and the sound analysis result to obtain a biodiversity prediction result specifically includes:

[0095] Obtaining the real biodiversity results of the target sea area, performing correlation analysis on the biometric identification results and the sound analysis results with the biodiversity results to obtain a first analysis result, and performing correlation analysis on the biometric identification results and the sound analysis results with the real biodiversity results to obtain a second analysis result;

[0096] Data screening is performed on the first analysis result and the second analysis result to obtain a biodiversity assessment factor, and biodiversity prediction is performed based on the biodiversity assessment factor to obtain a predicted biodiversity result.

[0097] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a multi-dimensional underwater biodiversity automated monitoring program, and when the multi-dimensional underwater biodiversity automated monitoring program is executed by a processor, the steps of the multi-dimensional underwater biodiversity automated monitoring method as described above are implemented.

[0098] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0099] Of course, those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing related hardware (such as a processor, a controller, etc.) through a computer program, and the program can be stored in a computer-readable storage medium that can be read by a computer, and the program can include the processes of the above-mentioned method embodiments when executed. The computer-readable storage medium can be a memory, a disk, an optical disk, etc.

[0100] It should be understood that the application of the present invention is not limited to the above examples. For ordinary technicians in this field, improvements or changes can be made based on the above description. All these improvements and changes should fall within the scope of protection of the claims attached to the present invention.

Claims

1. A multi-dimensional underwater biodiversity automated monitoring method, characterized in that: The multi-dimensional underwater biodiversity automated monitoring method comprises: Acquire environmental DNA data, underwater image data and underwater sound data of a target sea area, and pre-process the environmental DNA data, underwater image data and underwater sound data to obtain target environmental DNA data, target underwater image data and target underwater sound data; The created correlation network training model is trained according to the target environment DNA data, the target underwater image data and the target underwater sound data to obtain a target correlation analysis model; The training of the created correlation network training model according to the target environment DNA data, the target underwater image data and the target underwater sound data to obtain the target correlation analysis model specifically includes: Creating a relevant network training model, performing species diversity analysis on the target environmental DNA data to obtain a biodiversity result, using the biodiversity result as a validation set, and using the target underwater image data and the target underwater sound data as a training set, wherein the training set includes a plurality of training groups; Inputting a training group in the training set into the relevant network training model, performing species recognition processing on the target underwater image data to obtain a biological recognition result, and performing sound recognition analysis on the target underwater sound data to obtain a sound analysis result; Analyze the biometric recognition result and the sound analysis result to obtain a predicted biodiversity result, and calibrate model parameters of the relevant network training model according to the validation set and the predicted biodiversity result; Inputting the next training group into the relevant network training model until the training status of the relevant network training model meets the preset conditions, thereby obtaining a trained target correlation analysis model; The underwater image data and underwater sound data of the sea area to be tested are obtained, the underwater image data and underwater sound data to be tested are input into the target correlation analysis model, and the underwater biodiversity monitoring result is output.

2. The multi-dimensional underwater biodiversity automated monitoring method according to claim 1, characterized in that: The step of acquiring the environmental DNA data, underwater image data and underwater sound data of the target sea area, and preprocessing the environmental DNA data, the underwater image data and the underwater sound data to obtain the target environmental DNA data, the target underwater image data and the target underwater sound data specifically includes: Setting a sampling depth, performing multi-channel sampling on the biological DNA data of the target sea area according to the sampling depth to obtain environmental DNA data, performing image acquisition on the underwater organisms in the target sea area according to the sampling depth to obtain underwater image data, and performing sound acquisition on the target sea area according to the sampling depth to obtain underwater sound data; The environmental DNA data, the underwater image data and the underwater sound data are subjected to data review to obtain corresponding abnormal values, and the abnormal values ​​corresponding to the environmental DNA data, the underwater image data and the underwater sound data are corrected to obtain target environmental DNA data, target underwater image data and target underwater sound data, respectively. The abnormal values ​​include missing values, error values ​​and abnormal outliers.

3. The multi-dimensional underwater biodiversity automated monitoring method according to claim 1, characterized in that: The performing of species diversity analysis on the target environmental DNA data to obtain biodiversity results specifically includes: Acquire a target target region of the target environmental DNA data, and perform amplification processing on primers of the target target region to obtain an amplicon sequence; The amplicon sequence is subjected to high-throughput sequencing to obtain a sequencing result, and data analysis is performed on the sequencing result to obtain a biodiversity result, wherein the biodiversity result includes species composition and biodiversity information.

4. The multi-dimensional underwater biodiversity automated monitoring method according to claim 1, characterized in that: The performing species identification processing on the target underwater image data to obtain a biological identification result specifically includes: Performing image preprocessing on the target underwater image data to obtain a target image, wherein the image preprocessing includes color correction processing, defogging processing and filtering processing; Based on the trained image recognition model, the target image is subjected to feature extraction to obtain multiple feature information, all the feature information is fused to obtain the target feature, the organisms in the target image are marked according to the target feature to obtain the number of organisms, and the number of organisms is classified into species to obtain the biological recognition result.

5. The multi-dimensional underwater biodiversity automated monitoring method according to claim 1, characterized in that: The performing sound recognition analysis on the target underwater sound data to obtain a sound analysis result specifically includes: Performing high-frequency signal detection on the target underwater sound data according to a first preset signal detection threshold to obtain a target high-frequency signal, performing feature extraction on the target high-frequency signal to obtain a target high-frequency feature, and obtaining acoustic complexity and sound diversity according to the target high-frequency feature; Performing low-frequency signal detection on the target underwater sound data according to a second preset signal detection threshold to obtain a target low-frequency signal, performing feature extraction on the target low-frequency signal to obtain a target low-frequency feature, and obtaining sound richness and species type according to the target low-frequency feature; Calculating the sound pressure of the target underwater sound data to obtain the sound pressure intensity and the octave band sound pressure level; A sound analysis result is obtained according to the acoustic complexity, the sound diversity, the sound richness, the species type, the sound pressure intensity and the octave band sound pressure level.

6. The multi-dimensional underwater biodiversity automated monitoring method according to claim 1, characterized in that: The analyzing the biometric identification result and the sound analysis result to obtain a predicted biodiversity result specifically includes: Obtaining the real biodiversity results of the target sea area, performing correlation analysis on the biometric identification results and the sound analysis results with the biodiversity results to obtain a first analysis result, and performing correlation analysis on the biometric identification results and the sound analysis results with the real biodiversity results to obtain a second analysis result; Data screening is performed on the first analysis result and the second analysis result to obtain a biodiversity assessment factor, and biodiversity prediction is performed based on the biodiversity assessment factor to obtain a predicted biodiversity result.

7. A multi-dimensional underwater biodiversity automated monitoring system, characterized in that: The multi-dimensional underwater biodiversity automated monitoring system comprises: A data acquisition module is used to acquire environmental DNA data, underwater image data and underwater sound data of a target sea area, and pre-process the environmental DNA data, underwater image data and underwater sound data to obtain target environmental DNA data, target underwater image data and target underwater sound data; A data processing module, used for training the created correlation network training model according to the target environment DNA data, the target underwater image data and the target underwater sound data to obtain a target correlation analysis model; The training of the created correlation network training model according to the target environment DNA data, the target underwater image data and the target underwater sound data to obtain the target correlation analysis model specifically includes: Creating a relevant network training model, performing species diversity analysis on the target environmental DNA data to obtain a biodiversity result, using the biodiversity result as a validation set, and using the target underwater image data and the target underwater sound data as a training set, wherein the training set includes a plurality of training groups; Inputting a training group in the training set into the relevant network training model, performing species recognition processing on the target underwater image data to obtain a biological recognition result, and performing sound recognition analysis on the target underwater sound data to obtain a sound analysis result; Analyze the biometric recognition result and the sound analysis result to obtain a predicted biodiversity result, and calibrate model parameters of the relevant network training model according to the validation set and the predicted biodiversity result; Inputting the next training group into the relevant network training model until the training status of the relevant network training model meets the preset conditions, thereby obtaining a trained target correlation analysis model; The biological monitoring module is used to obtain underwater image data and underwater sound data of the sea area to be tested, input the underwater image data and underwater sound data to be tested into the target correlation analysis model, and output underwater biodiversity monitoring results.

8. A terminal, characterized in that: The terminal includes a memory, a processor, and a program stored in the memory and executable on the processor. When the program is executed by the processor, the steps of the multi-dimensional underwater biodiversity automated monitoring method as described in any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium, characterized in that: A computer program is stored thereon, wherein the computer-readable storage medium stores a multi-dimensional underwater biodiversity automated monitoring program, and when the multi-dimensional underwater biodiversity automated monitoring program is executed by a processor, the steps of the multi-dimensional underwater biodiversity automated monitoring method as claimed in any one of claims 1 to 6 are implemented.