Method and system for identifying species of marine biota

Abstract

The embodiment of the application discloses a kind of marine biological population species identification method and system, it is related to target identification technical field, including: in the case where wideband frequency modulation acoustic signal is launched to the early warning sea area of nuclear power water intake, receive the N frame target acoustic echo signal of target marine biological population;Target acoustic echo signal corresponds with wideband frequency modulation acoustic signal;N is positive integer;According to N frame target acoustic echo signal, determine the target frequency response curve of target marine biological population;Target frequency response curve is used to characterize the relative morphology of the volume backscattering intensity of the same space target biological population region formed by target marine biological population with frequency variation;According to target frequency response curve, determine the target species of target marine biological population.The application utilizes the volume scattering intensity difference of different biological population under different acoustic frequencies, extracts the frequency response curve of target as acoustic fingerprint, can accurately realize the species identification of nuclear power cold source disaster-causing biological population.

Description

Technical Field

[0001] This application relates to the field of target recognition technology, and in particular to a method and system for identifying marine biota species. Background Technology

[0002] Nuclear power plants require large amounts of seawater as a cooling source for their cooling systems. In recent years, due to changes in the marine ecological environment, micro-organisms such as jellyfish, shrimp, and fish have frequently proliferated in the water intake area of ​​nuclear power plants. These high-density organisms can be carried by ocean currents into the intake, clogging the intake facilities and threatening the safe operation of the nuclear power units. To ensure the safety of the nuclear power plant's cooling source, it is necessary to identify the species of marine organisms in the water intake area.

[0003] In related technologies, optical video surveillance or sonar monitoring technologies are commonly used to detect marine biota in the water intake area of ​​nuclear power plants. Optical video surveillance monitors biota by acquiring underwater images, but it is difficult to effectively identify the species of marine biota in complex aquatic environments. Sonar monitoring technology detects biota by emitting sound waves into the sea and receiving the echoes, but it is also difficult to identify the species of marine biota.

[0004] It is evident that current methods for detecting marine biota in the water intake area of ​​nuclear power plants present a challenge in identifying the species within the biota. Summary of the Invention

[0005] In view of this, one of the objectives of this application is to provide a method and system for identifying marine biota species, which can accurately identify marine biota species.

[0006] To achieve the above objectives, the technical solution of this application is implemented as follows: In a first aspect, embodiments of this application provide a method for identifying marine biota species, the method comprising: When transmitting broadband FM acoustic signals to the warning sea area of ​​the nuclear power plant's water intake, N frames of target acoustic echo signals of the target marine organisms are received; the target acoustic echo signals correspond to the broadband FM acoustic signals; N is a positive integer; Based on N frames of target acoustic echo signals, the target frequency response curve of the target marine biota is determined; the target frequency response curve is used to characterize the relative shape of the volume backscattering intensity of the target marine biota formation within the same spatial target biota region as a function of frequency. Based on the target frequency response curve, the target species of the target marine organism group are determined.

[0007] In one possible implementation, receiving N frames of target acoustic echo signals from the target marine biota includes: The target acoustic echo signal is received using an underwater acoustic transducer array.

[0008] In one possible implementation, the method further includes: Obtain marine detection information of the warning sea area; the marine detection information includes at least one of the following: first width, first distance, first mainstream direction, first environmental noise, and first echo signal-to-noise ratio; the first distance is the distance between the warning sea area and the target marine biota; Based on the marine exploration information, the underwater acoustic transducer array is adjusted so that the transmitting beam of the underwater acoustic transducer array covers the water intake section of the warning sea area, or so that the receiving beam of the underwater acoustic transducer array is directed towards the target biota area and the signal-to-noise ratio of the target biota area is improved.

[0009] In one possible implementation, before determining the target frequency response curve of the target marine biota based on N frames of target acoustic echo signals, the method further includes: N frames of target acoustic echo signals are preprocessed to obtain N frames of preprocessed target acoustic echo signals. The preprocessing includes at least one of matched filtering, time-varying gain and transmission loss compensation, and background reverberation removal. Matched filtering is used to improve the resolution and signal-to-noise ratio of the first distance. Time-varying gain and transmission loss compensation is used to compensate for the geometric diffusion attenuation and absorption attenuation of the target acoustic echo signal in seawater. Background reverberation removal is used to remove reverberation interference at fixed boundaries. The first distance is the distance between the warning sea area and the target marine biota. Based on N frames of target acoustic echo signals, determine the target frequency response curve of the N-frame target marine biota, including: Based on the N frames of preprocessed target acoustic echo signals, the target frequency response curve of the target marine biota is determined.

[0010] In one possible implementation, the target frequency response curve of the target marine biota is determined based on N frames of target acoustic echo signals, including: The N frames of target acoustic echo signals are converted into N corresponding target volume backscattering intensity matrices; the target volume backscattering intensity matrix varies with the first distance and frequency; the first distance is the distance between the warning sea area and the target marine organism group; Clustering the backscattering intensity matrices of N target volumes yields N frames of target biomes. Based on the target biome region in N frames, the target frequency response curve is determined.

[0011] In one possible implementation, before determining the target frequency response curve based on the N-frame target biome region, the method further includes: Obtain N first feature information corresponding to the target biome region in N frames; the first feature information includes at least one of the following: first centroid, first boundary, first area, first moving speed, and first cross-frequency continuity; Based on N first feature information, non-target interference is removed from the N frames of target biosphere region to obtain the N frames of target biosphere region after removal; non-target interference includes at least one of isolated noise point interference, fixed boundary point interference and ship bubble interference. Based on the target biome region in N frames, the target frequency response curve is determined, including: The target frequency response curve is determined based on the target biota region after N frames of rejection.

[0012] In one possible implementation, the target frequency response curve is determined based on the target biome region after N frames of rejection, including: For each frame of the target biomass region after removal, the volume backscattering intensity at each frequency point within its occupied space is averaged to obtain the scattering intensity at the first discrete frequency point of N frames. The target frequency response curve is determined based on the scattering intensity of the first discrete frequency point in N frames.

[0013] In one possible implementation, determining the target frequency response curve based on the scattering intensity of the first discrete frequency point in N frames includes: Perform trajectory association on the target biome region of N frames to identify M target biome regions belonging to the same trajectory; M is less than or equal to N and is a positive integer; The scattering intensities of the M first discrete frequency points corresponding to the M target biological community regions are weighted and averaged to obtain the scattering intensities of the second discrete frequency points after cross-frame averaging. The target frequency response curve is determined based on the scattering intensity at the second discrete frequency point.

[0014] In one possible implementation, determining the target frequency response curve based on the scattering intensity at the second discrete frequency point includes: The scattering intensity at the second discrete frequency point is smoothed and normalized to obtain the scattering intensity at the third discrete frequency point. The scattering intensity at the third discrete frequency point is resampled to the same frequency grid to obtain the target frequency response curve.

[0015] In one possible implementation, resampling the scattering intensity at the third discrete frequency point to the same frequency grid includes: Hermitian interpolation was used to resample the scattering intensity at the third discrete frequency point to the same frequency grid.

[0016] In one possible implementation, the target species of the target marine biota are determined based on the target frequency response curve, including: Obtain the P first similarity scores between the target frequency response curve and P reference continuous frequency response curves; P is a positive integer. Based on P first similarity scores, determine the first continuous frequency response curve from P reference continuous frequency response curves; The reference type corresponding to the first continuous frequency response curve is determined as the target type.

[0017] In one possible implementation, the target species of the target marine biota are determined based on the target frequency response curve, including: Extract the first feature vector of the target frequency response curve; the first feature vector includes at least one of the following: slope, peak value, curvature, monotonicity, bandwidth area, target biomass region variance, and cross-frame stability. A target deep learning classifier is used to process the first feature vector to obtain the target category. The target deep learning classifier includes an input layer, a hidden layer and an output layer connected in sequence. The hidden layer is used to extract local slope changes and peak combination features in the first feature vector. The output layer uses a normalized exponential function to output the category probability.

[0018] In one possible implementation, after determining the target species of the target marine biota based on the target frequency response curve, the method further includes: The target disaster risk level of the target marine biota is determined based on the biomass inverted from the target type and the backscatter intensity of the target volume. Output the target early warning information corresponding to the target's disaster risk level.

[0019] Secondly, embodiments of this application provide a marine biota species identification system, the system comprising: The receiving module is used to receive N frames of target acoustic echo signals from the target marine organisms when a broadband FM acoustic signal is transmitted to the warning sea area of ​​the nuclear power plant's water intake; the target acoustic echo signals correspond to the broadband FM acoustic signals; N is a positive integer; The first determining module is used to determine the target frequency response curve of the target marine biota based on N frames of target acoustic echo signals; the target frequency response curve is used to characterize the relative shape of the volume backscattering intensity of the target marine biota within the same spatial target biota region as a function of frequency. The second determination module is used to determine the target species of the target marine organism group based on the target frequency response curve.

[0020] The marine biota identification method provided in this application involves receiving N frames of target acoustic echo signals from a target marine biota after transmitting a broadband frequency-modulated acoustic signal to the warning sea area of ​​a nuclear power plant intake. Based on these N frames, a target frequency response curve for the target marine biota is determined. This curve characterizes the relative morphology of the volume backscattering intensity of the target marine biota within the same spatial target biota region as a function of frequency. Finally, the target species of the target marine biota is determined based on the target frequency response curve. This application utilizes the differences in volume backscattering intensity of different biota at different acoustic frequencies to extract the target's frequency response curve as an acoustic fingerprint, enabling accurate identification of biota that could cause damage from nuclear power plant cold sources. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. It should be understood that the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] The attached diagram is described below: Figure 1 One of the flowcharts for a method for identifying marine biota species provided in this application embodiment; Figure 2 A second schematic flowchart illustrating a method for identifying marine biota species provided in this application embodiment; Figure 3 A schematic diagram of the system structure involved in a method for identifying marine biota species provided in this application embodiment; Figure 4 A schematic diagram of the functional modules of a marine biota species identification system provided in this application embodiment; Figure 5 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of this application.

[0023] Explanation of reference numerals in the attached figures: 400. Marine biota species identification system; 410. Receiving module; 420. First Determined Module; 430. Second Determination Module; 501. Processor; 502. Memory; 503. Communication interface; 510. Bus. Detailed Implementation

[0024] The features and exemplary embodiments of various aspects of this application will be described in detail below. To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain this application and not to limit it. For those skilled in the art, this application can be implemented without some of these specific details. The following description of the embodiments is merely to provide a better understanding of this application by illustrating examples.

[0025] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes said element.

[0026] To facilitate a better understanding of the solutions in the embodiments of this application, the relevant technologies will be introduced first below.

[0027] Frequency Response Curve (FRC): A curve showing the relationship between the scattering intensity of a target at different acoustic frequencies and the frequency itself. In this application, it is used to describe the frequency response characteristics of a marine biological community per unit volume.

[0028] Time-Variable Gain (TVG): A technique that dynamically compensates for the gain of received echo signals based on changes in sound wave propagation distance and frequency. It is used to compensate for the geometric diffusion attenuation and absorption attenuation of sound waves in seawater.

[0029] Region of Interest (ROI) (can be used to refer to the target biological community region): The spatial region where the suspected biological community is located in the sonar detection results, used for subsequent feature extraction and analysis.

[0030] Constant False Alarm Rate (CFAR) detection is an adaptive threshold detection method that calculates the statistical characteristics of the local background by setting a training window around each range orientation cell, thereby forming an adaptive detection threshold.

[0031] Moving Target Indication (MTI) is a signal processing method that uses adjacent frame differential techniques to suppress reverberation interference at fixed boundaries and preserve the echo of moving targets.

[0032] Linear Frequency Modulation (LFM): A modulation method in which the signal frequency changes linearly with time, used in this application to generate a broadband acoustic transmission signal.

[0033] Analog-to-Digital Conversion (A / D): The process of converting analog signals into digital signals.

[0034] Piecewise Cubic Hermite Interpolating Polynomial (PCHIP) is an interpolation method that preserves the monotonicity and extremum positions of the original discrete points, without producing overshoot peaks that are not present in the original data.

[0035] The normalization function (Softmax) is a function that transforms any real vector into a probability distribution, and is often used in the output layer of neural network classifiers.

[0036] Target Strength (TS): The backscattering intensity of a single target in a specific orientation.

[0037] Volume Backscattering Strength (Sv): The backscattering intensity of a group of scattering particles per unit volume, used to describe the overall scattering characteristics of a biological population.

[0038] Source Level (SL): The sound pressure level of the sonar emitted signal.

[0039] Transmission Loss (TL): The energy attenuation of a sound wave as it propagates from its source to its target location.

[0040] Related technologies typically employ optical video surveillance or sonar monitoring to detect marine biota in the waters surrounding nuclear power plant intakes. Optical video surveillance monitors biota by acquiring underwater images, but it struggles to effectively identify species within complex aquatic environments. Sonar monitoring detects biota by emitting sound waves into the sea and receiving the echoes, but similarly, it also struggles to identify species within the marine ecosystem.

[0041] Based on this, in order to solve the related problems in the background art, embodiments of this application provide a method for identifying marine biota species, a system for identifying marine biota species, an electronic device, a computer-readable storage medium, and a computer program product.

[0042] The method for identifying marine biota species provided in the embodiments of this application will be described below. Please refer to... Figure 1 , Figure 1 This is one of the flowcharts illustrating a method for identifying marine biota species provided in an embodiment of this application, such as... Figure 1 The method shown can be applied to marine biota species identification systems or electronic devices in the following embodiments. In some embodiments, the electronic device includes a personal computer, server, mobile device, cloud computing platform, and supercomputer, etc. The following will describe its application to electronic devices, such as... Figure 1 The method shown specifically includes the following steps 110 to 130.

[0043] Step 110: While transmitting a broadband FM acoustic signal to the warning sea area of ​​the nuclear power plant's water intake, receive N frames of target acoustic echo signals from the target marine biota; the target acoustic echo signals correspond to the broadband FM acoustic signals; N is a positive integer.

[0044] Step 120: Determine the target frequency response curve of the target marine biota based on N frames of target acoustic echo signals; the target frequency response curve is used to characterize the relative shape of the volume backscattering intensity of the target marine biota formation within the same spatial target biota region as a function of frequency.

[0045] Step 130: Determine the target species of the target marine biota based on the target frequency response curve.

[0046] The marine biota identification method provided in this application involves receiving N frames of target acoustic echo signals from a target marine biota when a broadband frequency-modulated acoustic signal is transmitted to the warning sea area of ​​a nuclear power plant intake. Based on these N frames, a target frequency response curve is determined. This curve characterizes the relative shape of the volume backscattering intensity of the target marine biota within the same spatial target biota region as a function of frequency. Finally, the target species of the target marine biota is determined based on the frequency response curve. This application utilizes the differences in volume backscattering intensity of different biota at different acoustic frequencies to extract the target's frequency response curve as an acoustic fingerprint, enabling the identification of biota that could cause nuclear power plant cold source disasters. This solves the problem in related technologies where single-frequency sonar can only acquire target backscattering intensity but cannot distinguish target species.

[0047] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of related data must comply with relevant regulations. The acquisition, storage, use and processing of data in the technical solution of this application all comply with the relevant provisions of national laws and regulations.

[0048] In step 110, the electronic device can receive N frames of target acoustic echo signals of the target marine biota while transmitting broadband FM acoustic signals to the warning sea area of ​​the nuclear power plant's water intake.

[0049] The aforementioned nuclear power plant water intake refers to the inlet area of ​​the water intake channel in the cooling system of a nuclear power plant used for drawing seawater. The cooling system of a nuclear power plant requires the extraction of large amounts of seawater as a cold source, and the water intake is typically equipped with interception devices such as rotating filters to prevent marine life and floating debris from entering the cooling system.

[0050] The aforementioned warning sea area refers to a specific sea area in front of the nuclear power plant's water intake that requires continuous monitoring. The extent of this area can be determined based on factors such as the nuclear power plant's engineering layout, ocean current direction, and water intake flow rate, and typically covers a certain distance and width of sea area in front of the water intake canal.

[0051] The aforementioned broadband frequency-modulated acoustic signal refers to a frequency-modulated acoustic signal covering a relatively wide frequency band. Broadband signals can simultaneously acquire target scattering information at multiple frequency points, providing a data basis for subsequent extraction of frequency response curves. For example, the transmission frequency band used in this application embodiment is a 70kHz to 150kHz LFM broadband signal, with a pulse width set to 5ms, thus balancing detection range and range resolution.

[0052] The aforementioned target marine biota can refer to small marine organisms that may cause nuclear power plant cooling source disasters, including but not limited to jellyfish, shrimp, schools of fish, and large algae. These biota frequently erupt in the sea area near the nuclear power plant's intake, and can be carried by ocean currents into the intake channel, clogging the rotating filter and leading to the loss of cooling source.

[0053] The aforementioned target acoustic echo signal refers to the backscattered echo signal generated when a broadband frequency-modulated acoustic signal encounters a target marine biota. When the emitted sound wave encounters a marine biota, backscattering occurs, and the receiving device can acquire this echo signal. Because the biota has different scattering characteristics at different frequencies, the echo signal contains the target's frequency response information.

[0054] N is a positive integer representing the number of received echo signal frames. N frames of target acoustic echo signals are consecutive frames of echo data received at different times, which can be used for subsequent spatiotemporal averaging to reduce the impact of instantaneous noise and improve the stability of the frequency response curve. For example, N can be an integer between 10 and 30, and the specific value can be determined based on the environmental noise level of the warning sea area and the target's movement speed.

[0055] In one possible implementation, receiving N frames of target acoustic echo signals from a target marine biota includes: receiving N frames of target acoustic echo signals using an underwater acoustic transducer array.

[0056] In this embodiment, by using an underwater acoustic transducer array to receive echo signals, it is possible to achieve simultaneous reception of multiple channels, improve the spatial coverage and signal-to-noise ratio of the echo signals, and provide a data foundation for subsequent beamforming and target localization.

[0057] The aforementioned underwater acoustic transducer array can refer to a receiving array composed of several acoustic transducers, which can be used to convert received acoustic signals into electrical signals.

[0058] In some embodiments, the form of the underwater acoustic transducer array can be set according to the engineering conditions of the water intake, including but not limited to the following forms: a one-dimensional linear array arranged laterally along the bank or in front of the gate of the water intake channel, a two-dimensional surface array installed on the bottom of the channel or on the platform, or a distributed array composed of several broadband transducer nodes.

[0059] In some embodiments, the underwater acoustic transducer array includes at least one transmitting element and multiple receiving elements, or includes several transceiver elements, and the installation position, pointing angle and sensitivity parameters of each element are pre-calibrated and stored.

[0060] For example, in a nuclear power plant cold source early warning scenario, a one-dimensional linear array is used, which is horizontally arranged along the banks of both sides of the water intake channel. The linear array includes one transmitting element and 16 receiving elements. The transmitting element is placed on the center line of the channel, and the receiving elements are distributed at equal intervals on both sides of the channel. The installation depth of each element is determined according to the cross-sectional shape of the water intake to ensure that the beam can cover the entire water intake section.

[0061] In one possible implementation, the method further includes: Obtain marine detection information of the warning sea area; the marine detection information includes at least one of the following: first width, first distance, first mainstream direction, first environmental noise, and first echo signal-to-noise ratio; the first distance is the distance between the warning sea area and the target marine biota; Based on the marine exploration information, the underwater acoustic transducer array is adjusted so that the transmitting beam of the underwater acoustic transducer array covers the water intake section of the warning sea area, or so that the receiving beam of the underwater acoustic transducer array is directed towards the target biota area and the signal-to-noise ratio of the target biota area is improved.

[0062] In this embodiment of the application, by adaptively adjusting the underwater acoustic transducer array based on the marine detection information, the array can maintain optimal detection performance under different operating conditions, thereby improving the quality of the echo signal and the reliability of target detection.

[0063] The aforementioned marine detection information can be used to reflect the real-time status of the warning marine area. For example, it may include at least one of the following: first width, first distance, first mainstream direction, first environmental noise, and first echo signal-to-noise ratio.

[0064] Specifically: The first width can refer to the width of the water intake section of the warning sea area, that is, the lateral span of the sea area to be monitored; The first distance can refer to the distance between the warning sea area and the target marine organism group, which can be calculated from the arrival time of the echo signal; The primary current direction can indicate the main flow direction of ocean currents in the warning area, and can be used to determine the movement trend of biological communities and possible paths into the water intake. The first environmental noise can refer to the background noise level in the warning sea area, such as ship noise, wave noise, biological noise, etc. The first echo signal-to-noise ratio (SNR) refers to the ratio of the intensity of the target acoustic echo signal to the intensity of the background noise, and can be used to reflect the reliability of target detection.

[0065] The aforementioned adjustment of the underwater acoustic transducer array based on marine detection information refers to the ability of the array to adaptively adjust to factors such as the width of the warning sea area, target distance, main current direction, environmental noise, and echo signal-to-noise ratio during operation.

[0066] Specifically, when in search mode, the electronic equipment controls the transmission beam to be wider and cover the entire water intake section to achieve a comprehensive scan of the warning sea area; when a suspected target group is detected, the electronic equipment controls the receiving beam formation weight, transmission subarray or operating frequency band according to the target azimuth to make the beam point to the target ROI and improve the signal-to-noise ratio of the area; when a certain array element is blocked, malfunctions or affected by local strong noise, the electronic equipment removes the array element or reduces its weight, and the remaining array elements form an effective subarray to continue working.

[0067] For example, when the electronic device detects that the signal-to-noise ratio of the first echo is lower than a preset threshold, it can increase the transmission power or narrow the receiving beam width to improve the signal-to-noise ratio of the target biosphere region; when the electronic device detects that the first mainstream direction has changed, it can adjust the beam pointing angle so that the beam covers the path that the biosphere may pass through.

[0068] In one possible implementation, before determining the target frequency response curve of the target marine biota based on N frames of target acoustic echo signals, the method further includes: N frames of target acoustic echo signals are preprocessed to obtain N frames of preprocessed target acoustic echo signals. The preprocessing includes at least one of matched filtering, time-varying gain and transmission loss compensation, and background reverberation removal. Matched filtering is used to improve the resolution and signal-to-noise ratio of the first distance. Time-varying gain and transmission loss compensation is used to compensate for the geometric diffusion attenuation and absorption attenuation of the target acoustic echo signal in seawater. Background reverberation removal is used to remove reverberation interference at fixed boundaries. The first distance is the distance between the warning sea area and the target marine biota. Based on N frames of target acoustic echo signals, determine the target frequency response curve of the N-frame target marine biota, including: Based on the N frames of preprocessed target acoustic echo signals, the target frequency response curve of the target marine biota is determined.

[0069] In some embodiments, adaptive thresholding or MTI techniques can be used to remove reverberation interference at fixed boundaries.

[0070] In the embodiments of this application, by performing preprocessing on the echo signal such as matched filtering, time-varying gain and transmission loss compensation, and background reverberation removal, the distance resolution and signal-to-noise ratio of the echo signal can be effectively improved, fixed boundary interference can be eliminated, and the accuracy and reliability of subsequent frequency response curve extraction can be improved.

[0071] The aforementioned matched filtering process is used to improve the resolution and signal-to-noise ratio of the first range. Specifically, for the echo signal of the m-th receiving channel, the complex conjugate time-reversed signal of the transmitted signal is used as a matched filter to obtain the range-compressed echo in the frequency domain.

[0072] In the digital signal processing domain, the echo signal is rapidly convolved in the frequency domain using the complex conjugate of the transmitted signal to achieve pulse compression. Then, the time-domain signal is transformed into a time-frequency two-dimensional matrix through discrete Fourier transform. Delay compensation and weighted superposition are then performed on different receiving channels based on the array element positions to obtain a range-azimuth frequency data cube, further improving range resolution and signal-to-noise ratio.

[0073] For example, the transmitted signal is a 70kHz to 150kHz LFM broadband signal. After matched filtering, the distance resolution can reach the distance resolution corresponding to the pulse width, that is, the theoretical distance resolution corresponding to a 5ms pulse width, which is far superior to the distance resolution capability of a single-frequency sonar.

[0074] The aforementioned time-varying gain and transmission loss compensation processing is used to compensate for the geometric diffusion attenuation and absorption attenuation of the target acoustic echo signal in seawater.

[0075] Specifically, the target distance R is calculated based on the echo arrival time as follows: ; Where c is the local speed of sound.

[0076] Compensation is provided based on two-way propagation loss, propagation loss It can be represented as follows: ; in, This is the seawater absorption coefficient.

[0077] The electronic device generates a TVG gain that varies with distance and frequency based on the aforementioned propagation loss, and performs amplitude correction by combining the transmitting sound source level, receiving sensitivity, transducer directivity, and calibration constant, thereby compensating for the geometric diffusion attenuation and absorption attenuation of sound waves in seawater.

[0078] For example, when the target distance is 100 meters, the geometric diffusion attenuation is about 40 dB, and the seawater absorption attenuation is about several dB at 120 kHz. The electronic device generates the corresponding TVG gain curve according to the above propagation loss formula and performs joint compensation for the echo signal in terms of distance and frequency dimensions.

[0079] The aforementioned background reverberation removal process can be used to remove reverberation interference at fixed boundaries. Specifically, when there are no obvious biological targets or the targets are sparse, a background reverberation template is established using the median or low quantile of N consecutive frames of echoes. During real-time detection, the current volumetric scattering intensity matrix is ​​subtracted from the background template. Simultaneously, a training window is set around each range-azimuth cell to calculate the local background mean and standard deviation. A CFAR adaptive threshold is formed, where k is determined based on the desired false alarm rate. For fixed strong scattering objects such as dock boundaries, trash racks, and seabed, adjacent frame differential or moving target display technology can be combined to suppress fixed reverberation, retaining only bio-echoes that move with the water flow and appear continuously across multiple frequency points.

[0080] For example, for the concrete banks on both sides of the water intake channel, the position of the echo remains fixed between consecutive frames. This fixed reverberation can be suppressed by differential processing of adjacent frames, retaining only the echo signal of the biological community that moves with the water flow.

[0081] After preprocessing, the electronic equipment converts the preprocessed echo voltage signal into a volume backscattering intensity matrix that varies with distance and frequency. Specifically, the echo voltage after matched filtering and beamforming is converted into a received sound level, and then the volume backscattering intensity is calculated according to the sonar equation as follows:

[0082] ; in: SL represents the sound source level. G_T and G_R are the transmit and receive directional gains, respectively. V_cell is the sampling volume determined by the pulse length, beamwidth, and distance: C_cal is the system calibration constant.

[0083] This yields the features used for subsequent feature extraction. matrix.

[0084] In step 120, the electronic device determines the target frequency response curve of the target marine biota based on N frames of target acoustic echo signals.

[0085] It should be noted that the target frequency response curve describes the frequency response of the entire biological population per unit volume, rather than the instantaneous target intensity of a single individual.

[0086] In one possible implementation, the target frequency response curve of the target marine biota is determined based on N frames of target acoustic echo signals, including: The N frames of target acoustic echo signals are converted into N corresponding target volume backscattering intensity matrices; the target volume backscattering intensity matrix varies with the first distance and frequency; the first distance is the distance between the warning sea area and the target marine organism group; Clustering the backscattering intensity matrices of N target volumes yields N frames of target biomes. Based on the target biome region in N frames, the target frequency response curve is determined.

[0087] In this embodiment of the application, by converting the echo signal into a volume backscattering intensity matrix and performing clustering, the spatial region of the target biota can be accurately located from the complex echo data, providing a precise spatial range for the subsequent extraction of the frequency response curve.

[0088] The aforementioned target volume backscattering intensity matrix varies with the first distance and frequency, i.e. The matrix is ​​defined as follows: R is the distance, θ is the azimuth, and f is the frequency. Each frame of echo signal can be converted into an Sv matrix after preprocessing, and N frames of echoes correspond to N Sv matrices.

[0089] The above-mentioned clustering of backscattering intensity matrices of N target volumes can refer to clustering connected components in the Sv matrix after CFAR detection, where the units exceed the threshold and are continuous in distance, orientation, and time, to obtain candidate biosphere regions, i.e. target biosphere regions.

[0090] Connected component clustering is a processing method that aggregates spatially adjacent units that meet detection conditions into a single region. For example, in a frame of the Sv matrix, if multiple units exceeding the CFAR threshold are detected, and these units are concentrated within a range of 80 to 120 meters in the distance dimension, within a range of ±15 degrees directly in front in the orientation dimension, and appear consecutively in multiple frames in time, then these units are clustered into a target biome region.

[0091] In one possible implementation, before determining the target frequency response curve based on the N-frame target biome region, the method further includes: Obtain N first feature information corresponding to the target biome region in N frames; the first feature information includes at least one of the following: first centroid, first boundary, first area, first moving speed, and first cross-frequency continuity; Based on N first feature information, non-target interference is removed from the N frames of target biosphere region to obtain the N frames of target biosphere region after removal; non-target interference includes at least one of isolated noise point interference, fixed boundary point interference and ship bubble interference. Based on the target biome region in N frames, the target frequency response curve is determined, including: The target frequency response curve is determined based on the target biota region after N frames of rejection.

[0092] In this embodiment of the application, by removing non-target interference from the target biota region based on the first feature information, it is possible to effectively remove interference such as isolated noise points, fixed boundary points and ship bubbles, thereby improving the accuracy of subsequent frequency response curve extraction.

[0093] The aforementioned first feature information may include at least one of the following: first centroid, first boundary, first area, first moving speed, and first transfrequency continuity.

[0094] Specifically: The first centroid can refer to the coordinates of the center position of the target biota region in the azimuth space, which is used to characterize the spatial position of the biota. The first boundary can refer to the outer contour of the target biota region, used to define the spatial extent of the biota. The first area can refer to the projected area of ​​the target biota region on the azimuth plane, which is used to reflect the population scale of the biota. The first moving speed can refer to the spatial displacement rate of the target biosphere region between consecutive frames, which is used to determine the motion characteristics of the target. The first cross-frequency continuity can refer to the degree of consistency in the spatial distribution of the target biota region at different frequency points, and is used to distinguish between real biota echoes and narrowband noise interference.

[0095] The aforementioned non-target interference removal based on the first feature information can refer to removing interference regions that do not conform to the movement laws of biological groups based on features such as the centroid, boundary, area, movement speed, and cross-frequency continuity of the ROI.

[0096] Specifically, isolated noise points are typically small in area and their positions change abruptly between adjacent frames, while the positions of fixed boundary points remain unchanged between frames. Ship bubble interference usually exhibits high-speed movement and poor cross-frequency continuity. By setting appropriate feature discrimination conditions, the above-mentioned interference can be eliminated.

[0097] For example, when the first area of ​​a cluster region is less than a preset area threshold and the position change between adjacent frames exceeds a preset distance threshold, the region is determined to be an isolated noise point and is removed; when the first movement speed of a cluster region is close to zero and the position coincides with the known dock boundary or trash rack position, the region is determined to be a fixed boundary point and is removed.

[0098] In one possible implementation, the target frequency response curve is determined based on the target biome region after N frames of rejection, including: For each frame of the target biomass region after removal, the volume backscattering intensity at each frequency point within its occupied space is averaged to obtain the scattering intensity at the first discrete frequency point of N frames. The target frequency response curve is determined based on the scattering intensity of the first discrete frequency point in N frames.

[0099] In this embodiment of the application, by averaging the volume backscattering intensity of each frequency point within the ROI, discrete frequency point scattering intensity describing the overall scattering characteristics of the biological population can be obtained, reducing the influence of local bright scattering points on the curve shape, thereby improving the stability and reliability of the extracted frequency response curve.

[0100] The aforementioned averaging of volumetric backscattering intensity at each frequency point refers to averaging the Sv values ​​of all effective sampling units within the same ROI at each frequency point. Specifically, the Sv in dB form can be first converted to the linear energy domain for averaging, and then converted back to dB form, i.e.: ; Where n represents the effective sampling units within the ROI, and N represents the number of effective units.

[0101] In some embodiments, a truncated average or median average after removing the highest and lowest proportions of outliers can be used to reduce the influence of local bright scattering points on the curve shape.

[0102] The aforementioned first discrete frequency scattering intensity refers to the Sv value sequence at each frequency point obtained after averaging each frame of the ROI. Each frame corresponds to a set of discrete frequency scattering intensities. For example, for a certain frame of ROI, Sv values ​​at 20 frequency points are extracted in the frequency band from 70kHz to 150kHz. First, the dB value of each frequency point is converted to the linear energy domain, averaged over 100 effective sampling units within the ROI, and then the average value is converted back to dB form to obtain the 20 first discrete frequency scattering intensity values ​​for that frame.

[0103] In one possible implementation, determining the target frequency response curve based on the scattering intensity of the first discrete frequency point in N frames includes: Perform trajectory association on the target biome region of N frames to identify M target biome regions belonging to the same trajectory; M is less than or equal to N and is a positive integer; The scattering intensities of the M first discrete frequency points corresponding to the M target biological community regions are weighted and averaged to obtain the scattering intensities of the second discrete frequency points after cross-frame averaging. The target frequency response curve is determined based on the scattering intensity at the second discrete frequency point.

[0104] In the embodiments of this application, by performing trajectory association and weighted averaging on multiple consecutive frames of ROI of the same biological population, the influence of instantaneous noise can be further reduced, the spatiotemporal stability of the frequency response curve can be improved, and the obtained curve describes the frequency response characteristics of the biological population as a whole rather than the instantaneous scattering of a single individual.

[0105] The aforementioned trajectory association refers to matching and associating Regions of Interest (ROIs) belonging to the same biogroup in different frames to form continuous motion trajectories. Trajectory association can be achieved by comparing indicators such as centroid distance, area change rate, and consistency of motion direction between ROIs in adjacent frames. After trajectory association, M ROIs belonging to the same trajectory can be identified from N frames of ROIs, where M is less than or equal to N and M is a positive integer.

[0106] The aforementioned weighted average refers to the weighted summation of the scattering intensities at the first discrete frequency points corresponding to the M ROIs according to their respective weights. In some embodiments, the weights can be determined by the signal-to-noise ratio (SNR) at each frequency point, the area stability of the ROI, and the consistency of target motion. Specifically, frames with higher SNR, more stable area, and more consistent motion have greater weights.

[0107] The aforementioned second discrete frequency scattering intensity refers to the sequence of discrete frequency scattering intensity values ​​obtained after cross-frame weighted averaging. For example, if an electronic device performs trajectory association on 10 frames of ROI and determines that 8 frames belong to the same biological group trajectory, then the first discrete frequency scattering intensities of these 8 frames are weighted and averaged according to their respective signal-to-noise ratio and area stability weights to obtain a set of second discrete frequency scattering intensities.

[0108] In one possible implementation, determining the target frequency response curve based on the scattering intensity at the second discrete frequency point includes: The scattering intensity at the second discrete frequency point is smoothed and normalized to obtain the scattering intensity at the third discrete frequency point. The scattering intensity at the third discrete frequency point is resampled to the same frequency grid to obtain the target frequency response curve.

[0109] This embodiment eliminates the influence of curve amplitude differences under different measurement conditions and highlights the morphological characteristics of the curve by smoothing and normalizing the scattering intensity of discrete frequency points and resampling it to a unified frequency grid.

[0110] The aforementioned smoothing filter refers to filtering the scattering intensity at discrete frequency points to reduce the impact of noise on the curve shape. When the measured curve has high noise, a low-order Savitzky-Golay filter or a cubic smoothing spline with a regularization term can be used for smoothing before interpolation.

[0111] In some embodiments, the selection criteria for the smoothing parameter are: the fitting residual does not exceed the standard deviation of the measurement noise estimated by the blank water area or the calibration sample, and the number of newly added extreme points after smoothing does not exceed a preset threshold.

[0112] The normalization process described above refers to using the Sv value at a certain reference frequency as a reference to calculate the relative scattering intensity difference at each frequency point and construct the frequency response curve of the target biological population. In some embodiments, a reference frequency with a high signal-to-noise ratio and located in the middle of the transducer's effective bandwidth can be used as a reference, such as 120kHz, to calculate FRC(f_i) = Sv_ROI(f_i) - Sv_ROI(f_ref), which is the volume scattering intensity difference at each frequency point relative to 120kHz.

[0113] When the reference frequency is affected by narrowband noise, the mean value of Sv_ROI within the effective frequency band can also be used as a reference for zero-mean normalization.

[0114] The aforementioned third discrete frequency scattering intensity can refer to the sequence of discrete frequency scattering intensity values ​​after smoothing filtering and normalization.

[0115] The aforementioned resampling to the same frequency grid refers to resampling the scattering intensity values ​​at different discrete frequency points onto a unified frequency grid using an interpolation method to facilitate feature extraction. For example, discrete FRC points in the range of 70kHz to 150kHz are resampled onto a unified frequency grid with 1kHz intervals, resulting in frequency response curve data at 81 frequency points.

[0116] In one possible implementation, resampling the scattering intensity of the third discrete frequency point to the same frequency grid includes: resampling the scattering intensity of the third discrete frequency point to the same frequency grid using Hermite interpolation.

[0117] In this embodiment, by using Hermite conformal interpolation for resampling, the scattering intensity of discrete frequency points can be resampled to a unified frequency grid while maintaining the monotonicity and extreme value positions of the original data. This avoids introducing false peaks and ensures the authenticity and reliability of the frequency response curve.

[0118] The aforementioned Hermitian interpolation specifically refers to piecewise cubic Hermitian conformal interpolation (PCHIP), an interpolation method that preserves the monotonicity and extremum positions of the original discrete points. The selection criteria for PCHIP are to maintain the monotonicity and extremum positions of the original discrete points and to avoid generating overshoot peaks not present in the original data. When the signal-to-noise ratio at each frequency point is high, PCHIP conformal interpolation is used to preserve monotonicity and peak positions; if narrowband noise exists at local frequencies, low-order Savitzky-Golay filtering is first used for smoothing before PCHIP interpolation is applied.

[0119] In some embodiments, the selection criteria for the fitting or smoothing algorithm are: not introducing resonance peaks that do not exist in the original discrete points, not changing the position of the main peak frequency, and the fitting residual not exceeding the calibration noise level.

[0120] Considering that linear interpolation produces a broken line curve shape when the frequency point spacing is large, losing the smooth characteristics of the frequency response curve, and that ordinary cubic spline interpolation is prone to overshoot at points of drastic data point changes, introducing spurious peaks not present in the original data, thus interfering with species identification, the embodiments of this application employ PCHIP, which can maintain the smoothness of the curve while strictly adhering to the monotonicity and extreme value positions of the original data, ensuring that the interpolated curve faithfully reflects the true frequency response characteristics of the biological population.

[0121] In step 130, the electronic device determines the target species of the target marine organism group based on the target frequency response curve.

[0122] The aforementioned target species refer to the species categories of biological groups identified through feature matching and classification algorithms, including but not limited to jellyfish, krill, small fish with swim bladders, and large algae.

[0123] In one possible implementation, the target species of the target marine biota are determined based on the target frequency response curve, including: Obtain the P first similarity scores between the target frequency response curve and P reference continuous frequency response curves; P is a positive integer. Based on P first similarity scores, determine the first continuous frequency response curve from P reference continuous frequency response curves; The reference type corresponding to the first continuous frequency response curve is determined as the target type.

[0124] In this embodiment of the application, by performing similarity matching between the real-time target frequency response curve and the pre-established reference template curve, species identification can be achieved by utilizing the physical characteristics differences in the morphology of frequency response curves of different biological groups.

[0125] The aforementioned reference continuous frequency response curves may refer to template curves stored in a pre-established database of acoustic frequency responses of typical hazardous organisms from nuclear power plant cold sources. The database of acoustic frequency responses of typical hazardous organisms from nuclear power plant cold sources does not store the TS frequency curve of a single individual target, but rather the FRC template obtained by extracting the ROI, normalizing the Sv volume, normalizing the reference frequency, and averaging the species in a population state. This template includes the mean curve, variance curve, typical peak frequency range, slope range, and data acquisition condition information.

[0126] The aforementioned first similarity refers to the degree of similarity between the real-time extracted target frequency response curve and each reference template curve in the typical disaster-causing bioacoustic frequency response database of nuclear power cold sources. During matching, the real-time FRC and the database template are first resampled to the same frequency grid, and weights are set according to the signal-to-noise ratio and template variance of each frequency point. Then, the weighted Pearson correlation coefficient, weighted Euclidean distance, or dynamic time warping distance are calculated.

[0127] For the j-th type of template, the comprehensive similarity score Score_j can be calculated as follows: ; in, The weighted correlation coefficient is D_j, the normalized distance is P_j, the physical rule score is P_j, and w1, w2, and w3 are determined by the calibration samples.

[0128] The aforementioned first continuous frequency response curve can refer to the template curve among the P reference curves that has the highest similarity to the target frequency response curve. For example, the database stores reference FRC templates for three biological groups: jellyfish, krill, and small fish with swim bladders. Jellyfish, as fluid-like organisms, have acoustic impedances extremely close to seawater, and their frequency response curves are extremely flat between 70 and 150 kHz, with a slope almost zero. Krill, as elastic-shelled organisms, are typically in the transition region from Rayleigh scattering to geometric scattering in this frequency band, and their frequency response curves show a clear monotonically increasing characteristic with increasing frequency. Small fish with swim bladders, due to the presence of their swim bladders, exhibit a clear resonance peak in their frequency response curves near a specific frequency. The electronic device matches the target FRC extracted in real-time with each of the three templates, calculates the overall similarity, and if the krill template has the highest similarity and exceeds a preset threshold, the target species is identified as a krill group.

[0129] In one possible implementation, the target species of the target marine biota are determined based on the target frequency response curve, including: Extract the first feature vector of the target frequency response curve; the first feature vector includes at least one of the following: slope, peak value, curvature, monotonicity, bandwidth area, target biomass region variance, and cross-frame stability. A target deep learning classifier is used to process the first feature vector to obtain the target category. The target deep learning classifier includes an input layer, a hidden layer and an output layer connected in sequence. The hidden layer is used to extract local slope changes and peak combination features in the first feature vector. The output layer uses a normalized exponential function to output the category probability.

[0130] In this embodiment of the application, by extracting the multidimensional feature vector of the frequency response curve and using a deep learning classifier for classification, the morphological details of the curve can be fully utilized to improve the accuracy and robustness of category identification. At the same time, the risk of simple black-box classification is avoided through consistency verification.

[0131] The aforementioned first feature vector includes at least one of the following: slope, peak value, curvature, monotonicity, bandwidth area, target biomass region variance, and cross-frame stability. Specifically, the electronic device extracts features such as the full-band slope, segmented slope, main peak frequency, main peak prominence, main peak half-width at half-maximum, number of peaks, number of curvature changes, monotonicity coefficient, bandwidth area, and cross-frame stability of the real-time curve to form the first feature vector.

[0132] Specifically: The full-band slope can refer to the overall trend slope of the frequency response curve within the entire effective frequency band, and is used to distinguish between flat curves and monotonically increasing curves. The segmented slope can refer to the local slope of the curve in each sub-band after the effective frequency band is divided into several sub-bands, and is used to describe the shape change of the curve in different frequency bands. The main peak frequency can refer to the frequency corresponding to the peak with the largest amplitude in the frequency response curve, and is used to identify organisms with bubble resonators. The prominence of the main peak refers to the degree to which the amplitude of the main peak is prominent relative to the baseline of the curve, and is used to measure the significance of the resonance peak. The full width at half maximum (FWHM) of the main peak refers to the frequency width of the main peak at half its amplitude, and is used to describe the width and narrowness characteristics of the resonant peak. The number of peaks can refer to the total number of extreme points in a frequency response curve, and is used to distinguish between single-peak and multi-peak curves; The number of curvature changes can refer to the number of times the sign of the second derivative of the curve changes, and is used to describe the complexity of the curve's shape; The monotonicity coefficient refers to the degree to which a curve remains monotonic within its effective frequency band, and is used to distinguish between monotonically increasing curves and non-monotonic curves. The frequency band area refers to the integral area of ​​the frequency response curve within the effective frequency band and between the frequency axis, and is used to reflect the overall energy level of the curve. The variance of the target biome region can refer to the dispersion of the Sv value of each sampling unit within the ROI, which is used to describe the uniformity of the scattering characteristics within the biome. Cross-frame stability refers to the degree of morphological consistency of the frequency response curve across multiple consecutive frames, and is used to evaluate the temporal stability of a target.

[0133] The aforementioned deep learning classifier comprises an input layer, a hidden layer, and an output layer connected in sequence. The classifier first standardizes the first feature vector using the mean and standard deviation of the training set, and then inputs it into the neural network classifier. The input layer consists of the FRC (Functional Ratio) on a fixed-frequency grid and the aforementioned physical features. The hidden layer extracts local slope variations and peak combination features. The output layer uses the Softmax function to output the probability of each species category.

[0134] It should be noted that the classification results also need to be consistent with the curve cross-correlation score and the physical rule score to avoid simple black-box classification. When the scores of the three categories point to the same category or the comprehensive score is the highest and exceeds the threshold, the classification result with the highest confidence is output; otherwise, a suspected mixed biological population can be output and manual review or continued tracking can be prompted.

[0135] For example, the electronic device extracts features from the target FRC, such as a full-band slope of 0.02dB / kHz, a main peak frequency of 95kHz, a main peak prominence of 8dB, a peak count of 1, and a monotonicity coefficient of 0.3, to form a first feature vector. This feature vector is then input into a pre-trained deep learning classifier for the target. The classifier outputs a probability of 0.05 for jellyfish, 0.10 for krill, and 0.82 for small fish with swim bladders. Simultaneously, the curve cross-correlation score and the physical rule score both point to small fish with swim bladders. Since the scores for the three categories are consistent, the target species is determined to be a group of small fish with swim bladders.

[0136] In one possible implementation, after determining the target species of the target marine biota based on the target frequency response curve, the method further includes: The target disaster risk level of the target marine biota is determined based on the biomass inverted from the target type and the backscatter intensity of the target volume. Output the target early warning information corresponding to the target's disaster risk level.

[0137] In this embodiment of the application, by determining the disaster risk level based on the identified species and the inverted biomass and outputting early warning information, nuclear power plants can take precise preventive measures against different types of disaster-causing organisms, thereby improving the efficiency of nuclear power emergency response.

[0138] The aforementioned target hazard risk level refers to a comprehensive assessment of the threat posed by the current biological population to the safety of nuclear power plant cold sources, based on the identified biological species and the biomass derived from their volumetric scattering intensity. Different biological populations possess different hazard characteristics. For example, jellyfish populations are soft and easily deformable, readily clogging filter mesh; high-density shrimp populations can surge into the water intake in a short period; and small fish with swim bladders may carry air bubbles, causing additional interference to acoustic equipment. Therefore, based on the specific species identified and their biomass, the hazard risk can be classified into different levels.

[0139] The aforementioned target early warning information may refer to classified early warning signals generated based on the target's disaster risk level, including biomes species labels, current location distance and orientation, movement trajectory, biomass estimation, and suggested countermeasures.

[0140] In some embodiments, target warning information can be sent to the nuclear power plant's central control room or distributed control system to guide the on-site deployment of appropriate interception and salvage equipment in advance.

[0141] For example, the electronic device identifies the target species as a group of krill and calculates a biomass density of 500 shrimp per cubic meter based on the Sv value. Combining this with the destructive characteristics of the krill group, the device classifies the target as having a high risk level and sends an early warning message to the nuclear power plant's control room. This message includes a krill group tag, a distance of 120 meters, a forward orientation, a movement speed of 0.3 meters per second, and a recommendation to deploy a fine-mesh net in advance. Upon receiving the warning message, the control room can proactively activate retrieval equipment or adjust the water intake flow rate to avoid resource waste and delays caused by indiscriminate responses.

[0142] It should be noted that the various embodiments described in this application can be combined with each other or implemented individually without conflict, and this application does not limit this.

[0143] To facilitate understanding of the above embodiments, a specific embodiment will be used as an example below, which includes the following steps: S1, Broadband Acoustic Signal Transmission and Echo Reception. This involves transmitting a broadband frequency-modulated acoustic signal to the warning sea area near the nuclear power plant's water intake, covering a wide frequency band; and using an underwater acoustic transducer array to receive the acoustic echo signal from the target biota.

[0144] S2, echo signal preprocessing and volume scattering intensity calculation. This corresponds to the implementation process of preprocessing the N frames of the target acoustic echo signal in the aforementioned embodiment.

[0145] S3, Target Frequency Response Curve Extraction. This corresponds to the process in the aforementioned embodiment of determining the target frequency response curve of the target marine biota based on N frames of target acoustic echo signals.

[0146] S4, Species identification and matching based on prior database. This corresponds to the process in the aforementioned embodiments of determining the target species of the target marine biota based on the target frequency response curve.

[0147] S5, Disaster Risk Level Assessment and Early Warning Output. This refers to the implementation process of outputting target early warning information corresponding to the target disaster risk level in the aforementioned embodiments.

[0148] Please refer to the implementation process diagrams of S1 to S5 above. Figure 2 , Figure 2 This is a second schematic flowchart of a method for identifying marine biota species provided in an embodiment of this application.

[0149] In addition, please see Figure 3 , Figure 3 This is a schematic diagram of the system structure involved in a marine biota species identification method provided in an embodiment of this application. Figure 3 The system architecture includes: Broadband acoustic transceiver module: includes a broadband waveform signal generator, a power amplifier, and a broadband underwater acoustic transducer; High-speed data acquisition and digital signal processing module (i.e., high-speed AD acquisition unit): realizes high sampling rate A / D conversion and low-level signal processing such as matched filtering in S2 above; Feature extraction and intelligent classification server: runs the algorithm software of S3 and S4 mentioned above, and has a built-in acoustic feature database.

[0150] Early warning and control terminal (i.e., early warning terminal): used for three-dimensional visualization of the location, movement trajectory, species and risk level of biological communities.

[0151] Corresponding to the above method embodiments, this application also provides a marine biota species identification system. Please refer to [link to relevant documentation]. Figure 4 , Figure 4This is a functional module diagram of a marine biota species identification system provided in an embodiment of this application. The marine biota species identification system 400 includes: Receiver module 410 is used to receive N frames of target acoustic echo signals of the target marine organisms when a broadband frequency-modulated acoustic signal is transmitted to the warning sea area of ​​the nuclear power plant water intake; the target acoustic echo signals correspond to the broadband frequency-modulated acoustic signals; N is a positive integer; The first determining module 420 is used to determine the target frequency response curve of the target marine biota based on N frames of target acoustic echo signals; the target frequency response curve is used to characterize the relative shape of the volume backscattering intensity of the target marine biota within the same spatial target biota region as a function of frequency. The second determining module 430 is used to determine the target species of the target marine organism group based on the target frequency response curve.

[0152] The marine biota species identification system provided in this application embodiment can achieve, for example... Figure 1 The various processes implemented in the Chinese method embodiments can achieve similar or the same technical effects, and will not be described again here to avoid repetition.

[0153] In addition to the methods and systems provided in the above embodiments, this application also provides an electronic device, please refer to [link to previous document]. Figure 5 , Figure 5 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of this application.

[0154] The electronic device may include a processor 501 and a memory 502 storing computer program instructions.

[0155] Specifically, the processor 501 may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0156] Memory 502 may include mass storage for data or instructions. For example, and not limitingly, memory 502 may include a hard disk drive (HDD), floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 502 may include removable or non-removable (or fixed) media. Where appropriate, memory 502 may be internal or external to the integrated gateway disaster recovery device. In a particular embodiment, memory 502 is non-volatile solid-state memory.

[0157] In some embodiments, memory 502 may include read-only memory (ROM), random access memory (RAM), disk storage media device, optical storage media device, flash memory device, electrical, optical, or other physical / tangible memory storage device. Therefore, generally, memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described in the methods provided according to embodiments of this application.

[0158] The processor 501 implements the method provided in the above embodiments by reading and executing computer program instructions stored in the memory 502.

[0159] In one example, the electronic device may also include a communication interface 503 and a bus 510. The processor 501, memory 502, and communication interface 503 are connected via the bus 510 and communicate with each other.

[0160] The communication interface 503 is mainly used to realize communication between various modules, devices, units and / or equipment in the embodiments of this application.

[0161] Bus 510 includes hardware, software, or both, that couples components of an electronic device together. For example, and not limitingly, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, bus 510 may include one or more buses. Although specific buses are described and illustrated in embodiments of this application, this application contemplates any suitable bus or interconnect.

[0162] Furthermore, in conjunction with the methods provided in the above embodiments, this application embodiment can be implemented using a computer-readable storage medium. This computer-readable storage medium stores computer program instructions; when executed by a processor, these computer program instructions implement any of the methods in the above embodiments.

[0163] Furthermore, in conjunction with the methods provided in the above embodiments, this application embodiment can provide a computer program product to implement the methods. This program product is stored in a storage medium and executed by at least one processor to implement the various processes of the embodiments of the methods provided in the above embodiments, achieving similar or identical technical effects. To avoid repetition, further details are omitted here.

[0164] It should be clarified that this application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of this application is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of this application.

[0165] The functional blocks shown in the above-described structural diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this application are programs or code segments used to perform the required tasks. Programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried on a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Code segments can be downloaded via computer networks such as the Internet, intranets, etc.

[0166] It should also be noted that the exemplary embodiments mentioned in this application describe methods or systems based on a series of steps or apparatus. However, this application is not limited to the order of the above steps; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0167] The aspects of this disclosure have been described above with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It should be understood that each block in the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that these instructions, executable via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions / actions specified in one or more blocks of the flowchart illustrations and / or block diagrams. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field-programmable logic circuit. It is also understood that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can also be implemented by special-purpose hardware performing the specified functions or actions, or can be implemented by a combination of special-purpose hardware and computer instructions.

[0168] The above description is merely a specific implementation of this application. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the protection scope of this application.

Claims

1. A method for identifying marine biota species, characterized in that, include: When transmitting broadband FM acoustic signals to the warning sea area of ​​the nuclear power plant's water intake, receive N frames of target acoustic echo signals from the target marine organisms; The target acoustic echo signal corresponds to the broadband frequency-modulated acoustic signal; N is a positive integer; Based on the target acoustic echo signals of the N frames, determine the target frequency response curve of the target marine organism community; The target frequency response curve is used to characterize the relative shape of the volume backscattering intensity of the target marine biota within the same spatial target biota region as a function of frequency; Based on the target frequency response curve, the target species of the target marine biota are determined.

2. The method according to claim 1, characterized in that, The received N frames of target acoustic echo signals from the target marine biota include: The target acoustic echo signal is received using an underwater acoustic transducer array, which receives N frames.

3. The method according to claim 2, characterized in that, The method further includes: Obtain marine detection information of the warning sea area; the marine detection information includes at least one of a first width, a first distance, a first current direction, a first environmental noise, and a first echo signal-to-noise ratio; the first distance is the distance between the warning sea area and the target marine biota; Based on the marine area detection information, the underwater acoustic transducer array is adjusted so that the transmitting beam of the underwater acoustic transducer array covers the water intake section of the warning marine area, or so that the receiving beam of the underwater acoustic transducer array is directed towards the target biota area and the signal-to-noise ratio of the target biota area is improved.

4. The method according to claim 1, characterized in that, Before determining the target frequency response curve of the target marine biota based on the N frames of the target acoustic echo signal, the method further includes: The target acoustic echo signals of N frames are preprocessed to obtain N frames of preprocessed target acoustic echo signals. The preprocessing includes at least one of matched filtering, time-varying gain and transmission loss compensation, and background reverberation removal. The matched filtering is used to improve the resolution and signal-to-noise ratio at a first distance. The time-varying gain and transmission loss compensation is used to compensate for the geometric diffusion attenuation and absorption attenuation of the target acoustic echo signals in seawater. The background reverberation removal is used to remove reverberation interference at fixed boundaries. The first distance is the distance between the warning sea area and the target marine biota. The step of determining the target frequency response curve of the target marine biota based on the N frames of target acoustic echo signals includes: Based on the preprocessed target acoustic echo signal of N frames, the target frequency response curve of the target marine biota is determined.

5. The method according to claim 1, characterized in that, The step of determining the target frequency response curve of the target marine biota based on N frames of the target acoustic echo signal includes: The N frames of target acoustic echo signals are converted into N corresponding target volume backscattering intensity matrices; the target volume backscattering intensity matrix varies with a first distance and frequency; the first distance is the distance between the warning sea area and the target marine biota; Clustering the N target volume backscattering intensity matrices yields N frames of target biomes. Based on the target biota region in N frames, the target frequency response curve is determined.

6. The method according to claim 5, characterized in that, Before determining the target frequency response curve based on the target biome region using N frames, the method further includes: Obtain N first feature information corresponding to the target biome region in N frames; the first feature information includes at least one of the following: first centroid, first boundary, first area, first moving speed, and first cross-frequency continuity; Based on N first feature information, non-target interference is removed from the target biosphere region in N frames to obtain the target biosphere region after removal in N frames; the non-target interference includes at least one of isolated noise point interference, fixed boundary point interference and ship bubble interference. The determination of the target frequency response curve based on the target biome region in N frames includes: Based on the N frames of the removed target biomes, the target frequency response curve is determined.

7. The method as described in claim 6, characterized in that, The determination of the target frequency response curve based on the removed target biome region of N frames includes: For each frame of the target biomass region after removal, the volume backscattering intensity at each frequency point within its occupied space is averaged to obtain the first discrete frequency point scattering intensity of N frames. The target frequency response curve is determined based on the scattering intensity of the first discrete frequency point in N frames.

8. The method according to claim 7, characterized in that, The step of determining the target frequency response curve based on the scattering intensity of the first discrete frequency point in N frames includes: Trajectory association is performed on the target biome regions in N frames to identify M target biome regions belonging to the same trajectory; M is less than or equal to N and is a positive integer; The scattering intensities of the M first discrete frequency points corresponding to the M target biological community regions are weighted and averaged to obtain the scattering intensities of the second discrete frequency points after cross-frame averaging. The target frequency response curve is determined based on the scattering intensity at the second discrete frequency point.

9. The method according to claim 8, characterized in that, Determining the target frequency response curve based on the scattering intensity of the second discrete frequency point includes: The scattering intensity at the second discrete frequency point is smoothed, filtered, and normalized to obtain the scattering intensity at the third discrete frequency point. The scattering intensity at the third discrete frequency point is resampled to the same frequency grid to obtain the target frequency response curve.

10. The method according to claim 9, characterized in that, The step of resampling the scattering intensity of the third discrete frequency point to the same frequency grid includes: Hermitian interpolation is used to resample the scattering intensity of the third discrete frequency point to the same frequency grid.

11. The method according to claim 1, characterized in that, The determination of the target species of the target marine biota based on the target frequency response curve includes: Obtain P first similarity scores between the target frequency response curve and P reference continuous frequency response curves; where P is a positive integer. Based on P first similarities, a first continuous frequency response curve is determined from P reference continuous frequency response curves; The reference type corresponding to the first continuous frequency response curve is determined as the target type.

12. The method according to claim 1, characterized in that, The determination of the target species of the target marine biota based on the target frequency response curve includes: Extract the first feature vector of the target frequency response curve; the first feature vector includes at least one of the following: slope, peak value, curvature, monotonicity, bandwidth area, target biomass region variance, and cross-frame stability. The first feature vector is processed using a target deep learning classifier to obtain the target category; the target deep learning classifier includes an input layer, a hidden layer, and an output layer connected in sequence; the hidden layer is used to extract local slope changes and peak combination features in the first feature vector; the output layer outputs the category probability using a normalized exponential function.

13. The method according to claim 1, characterized in that, After determining the target species of the target marine biota based on the target frequency response curve, the method further includes: Based on the biomass retrieved from the backscattering intensity of the target type and its target volume, the target disaster risk level of the target marine biota is determined. Output the target early warning information corresponding to the disaster risk level of the target.

14. A marine biota species identification system, characterized in that, The system includes: The receiving module is used to receive N frames of target acoustic echo signals of the target marine organisms when a broadband frequency-modulated acoustic signal is transmitted to the warning sea area of ​​the nuclear power plant's water intake; the target acoustic echo signals correspond to the broadband frequency-modulated acoustic signals; N is a positive integer; The first determining module is used to determine the target frequency response curve of the target marine biota based on the N frames of target acoustic echo signals; the target frequency response curve is used to characterize the relative shape of the volume backscattering intensity of the target marine biota within the same spatial target biota region as a function of frequency; The second determining module is used to determine the target species of the target marine biota based on the target frequency response curve.

Images