CN117392527A

CN117392527A - High-precision underwater target classification detection method and model building method thereof

Info

Publication number: CN117392527A
Application number: CN202311685185.XA
Authority: CN
Inventors: 王柘; 郑冰; 张沁悦; 李继哲; 赵一萌; 张赛男
Original assignee: Ocean University of China
Current assignee: Ocean University of China
Priority date: 2023-12-11
Filing date: 2023-12-11
Publication date: 2024-01-12
Anticipated expiration: 2043-12-11
Also published as: CN117392527B

Abstract

The invention provides a high-precision underwater target classification detection method and a model building method thereof, which belong to the technical field of underwater image data processing and analysis, and adopt self-adaptive empirical mode decomposition (BEMD) to perform feature extraction on input image data to be detected, and extract characteristics such as a hurst index, texture intensity and the like from a residual component sequence generated by BEMD decomposition. And then, establishing a fusion mechanism through multi-dimensional extraction features to form fusion feature data. And finally, establishing a k-means clustering algorithm model based on the fusion data, and carrying out classified target detection on the data to be detected by adopting the model, thereby obtaining a final detection result. The method of the invention improves the underwater target detection effect by combining a plurality of factors such as adaptivity, feature extraction, k-means feature clustering and the like, and is especially suitable for complex underwater scenes.

Description

High-precision underwater target classification detection method and model building method thereof

Technical Field

The invention belongs to the technical field of underwater image data processing and analysis, and particularly relates to a high-precision underwater target classification detection method and a model building method thereof.

Background

The underwater ecological system has important roles in ecology and economy, and comprises valuable resources such as fish, coral reefs and the like, so that the underwater ecological system needs to be studied deeply and effectively protected. Technicians mainly pay attention to the fields of fish identification, quantity estimation, coral reef research, underwater cultural heritage protection and the like so as to know the health condition of an ecological system and the maintenance of the cultural heritage. However, underwater target detection faces many challenges, including light transmission problems, background interference, target occlusion, and shadows, and so on, and thus techniques such as feature extraction, machine learning, and deep learning are continually being developed to improve accuracy and efficiency. Therefore, the development of the underwater target detection technology aims to provide powerful support for the protection and management of marine resources and promote the sustainable development of a marine ecosystem.

Existing underwater target classification detection techniques mainly cover the following three types: optical imaging technology, sonar imaging technology, and underwater laser radar (LIDAR) technology.

The existing underwater target classification detection technology mainly comprises the following two types: traditional detection methods and deep learning methods.

The traditional detection method comprises the following steps: the traditional detection method relates to a technology based on feature extraction and model matching, such as Adaboost, multi-frame triplet mode, average movement segmentation, normalized cross correction and the like, and performs target classification detection and tracking based on features such as color, shape, texture and the like. However, these methods face challenges such as light transmission problems, background camouflage, target occlusion, and shadows, limiting their accuracy and applicability, and conventional methods have limited accuracy and applicability in complex underwater complex environments.

The deep learning method comprises the following steps: in particular Convolutional Neural Networks (CNNs) have shown potential in the detection of underwater target classification. The deep learning method utilizes a large-scale data set and a pre-training model to automatically extract the characteristics and obtain good results. Some deep learning methods include pooling based on spatial pyramids, pre-training models (e.g., VGGNet), automated CNN methods, and the like. However, the deep learning method also has the problems of dependence on a large amount of training data, high computational complexity, high requirement on hardware resources and the like.

In general, existing techniques for classifying underwater targets have drawbacks in overcoming light transmission problems, background camouflage, computational complexity, data dependence, and the like, and further solving these problems is needed to improve accuracy and applicability of underwater target detection.

Disclosure of Invention

In view of the above problems, the first aspect of the present invention provides a method for building a high-precision underwater target classification detection model, including the following steps:

s1, acquiring an underwater target detection image datasetAnd +.>Preprocessing, wherein the preprocessing comprises corresponding label labeling;

s2, based on a two-dimensional empirical mode decomposition strategy and a model BEMD, the data set preprocessed in S1 is subjected toPerforming intrinsic mode function decomposition (IMF), and extracting main frequency domain characteristic information from image data information;

s3, aiming at residual components of BEMD decompositionExtracting features including Hurst index->Computation and texture intensity->Estimating; and integrating the multidimensional information through a feature fusion mechanism to obtain fusion feature data ++>；

S4, establishing a k-means clustering algorithm model, and adopting fusion characteristic dataAnd performing model training and testing to obtain the underwater target classification detection model.

Preferably, the S1 specifically is:

collecting video data of an underwater object to be detected, and segmenting the collected image into pixelsShort video data +.2 s long>；

At uniform time intervals, each segmentCut into image frame data of sequence length 50；

For image frame dataCorresponding label marking is carried out on the target to be detected, and the result is used as a group of data +.>；

Repeatedly executing the above processes, and integrating the data processing results to obtain a data set。

Preferably, the S2 specifically is:

for all image frame data in each set of sequence dataBEMD decomposition is performed, and the input of BEMD is defined as:

(1)

wherein,for the image sequence to be decomposed, < > for>And->Pixel coordinates in the image;

thus, a set of dataAt the same time as input to obtain a sequenceAnd each coordinate pixel corresponds to the sequence +.>WhereinThe method comprises the steps of carrying out a first treatment on the surface of the Calculating an envelope maximum value +.for each pixel coordinate for the corresponding sequence>And envelope minimumEnvelope average +.>The calculation of (2) is as shown in formula (2):

(2)

will beAs a result of the primary decomposition of BEMD +.>And the residual component calculation is obtained as shown in formula (3):

(3)

judgingIf the decomposition termination condition is not satisfied, the residual component is continuously used as a new input signal, and the characteristic decomposition is repeatedly performed; on the contrary, if the termination condition is satisfied, the decomposition is stopped to finally obtain the total +.>An IMF component and a residual component, wherein +.>Indicating that +.>Sub-feature decomposition.

Preferably, the decomposition termination condition is specifically:

for the firstResidual results of the sub-feature decomposition->Definitions->The method comprises the following steps:

(4)

based on the calculation method, the decomposition termination condition is set asThe method comprises the steps of carrying out a first treatment on the surface of the The final image data is decomposed into:

(5)。

preferably, the residual component for BEMD decomposition in S3Extracting features, wherein the Hurst index +.>The acquisition process comprises the following steps:

the Hurst index Hurst is used to measure the extent of variation of the fluctuation range of the sequence data with time span and indirectly reflects the long memory of the sequence data, as shown in formula (6):

(6)

wherein,is the number of sequence data; />Is->The variation range of the observation points; />Is the standard deviation of these points, therefore +.>Indicate use +.>For->Performing standardization; />Is constant, & lt>Is Hurst index, andthe method comprises the steps of carrying out a first treatment on the surface of the When->When there is no correlation in the sequence data; when->When the sequence data has long memory; />It indicates that the data has strong volatility;

thus, the first and second substrates are bonded together,the calculation of (c) is represented by formula (7):

(7)

obtaining Hurst index by fitting linear function to input data distributionAnd at this time the input data is a residual component。

Preferably, the texture strength in S3The acquisition process and the characteristic data fusion mechanism of the system comprise:

texture intensity is defined as the logarithm of covariance of IMF component of BEMD decomposition, i.e.:

(8)

wherein,representing a covariance extraction operation;

will Hurst indexAnd texture intensity->Feature integration is performed, namely:

(9)

wherein a and b are constant parameters that keep K positive; the fusion feature policy is defined as:

(10)

wherein,is +.>An IMF component->For the corresponding fusion characteristic data, therefore +_for each set of initial data>Executing the feature extraction and fusion strategy to finally obtain a fused frequency domain feature data set。

Preferably, the training process of the k-means clustering algorithm model in the step S4 is as follows:

taking the fusion characteristic data in the S4 as an input characteristic vector of a target to be detected, and obtaining the total category number to be detected according to the tag categories in the data set; presetting the category number of the target to be detected asI.e. the number of preset clusters of the k-means algorithm is +.>Performing k-means unsupervised learning training based on the fusion characteristic data set to finally obtain +.>Personal clustering centerThe feature vector corresponding to the cluster center is a representative feature vector of such a target.

The invention provides a high-precision underwater target classification detection method, which is characterized by comprising the following steps of:

step 1, acquiring image sequence data to be detected;

step 2, based on a two-dimensional empirical mode decomposition strategy and a model BEMD, performing intrinsic mode function decomposition IMF on the image data in the step 1, and extracting main frequency domain characteristic information from the image data information;

step 3, residual component for BEMD decompositionExtracting features including Hurst index->Computation and texture intensity->Estimating; and integrating the multidimensional information through a feature fusion mechanism to obtain fusion feature data ++>；

Step 4, fusing the characteristic dataInputting the characteristic classification detection model into the underwater target classification detection model constructed by the construction method according to the first aspect; the characteristic classification detection process comprises the following steps: calculate->Vector and->Euclidean distance of each cluster center is shown in formula 11:

(11)

wherein,for the feature data dimension, for the input +.>Feature data and->The distance of the feature vector of the cluster center point is finally obtained to obtain a group of distance features +.>If the input feature data has the minimum distance with the clustering center feature, the sample to be detected has the maximum detection probability relative to the target class, and the class is used as a corresponding target detection result, namely:

(12)

wherein,class labels corresponding to the cluster centers +.>The final detection result is the target;

and 5, outputting a target classification detection result.

The third aspect of the present invention also provides a high-precision underwater object classification detection apparatus, the apparatus comprising at least one processor and at least one memory, the processor and memory being coupled; a computer-implemented program of the underwater target classification detection model constructed by the construction method according to the first aspect is stored in the memory; when the processor executes the computer executing program stored in the memory, the processor can be caused to execute a high-precision underwater target classification detection method.

The fourth aspect of the present invention also provides a computer-readable storage medium characterized in that: the computer readable storage medium stores a computer execution program of the underwater target classification detection model constructed by the construction method according to the first aspect, and when the computer execution program is executed by a processor, the processor can be caused to execute a high-precision underwater target classification detection method.

Compared with the prior art, the invention has the following beneficial effects:

(1) Adaptivity: the BEMD method has strong self-adaptability to different data signals and can adapt to different underwater scenes, so that the dependence of manual feature extraction is reduced.

(2) Higher detection accuracy: by introducing BEMD features and a multidimensional feature fusion mechanism, the accuracy of underwater target detection is improved.

(3) Tracking of time guidance: the method is helpful for tracking the target object, reduces the error rate and is particularly good in complex underwater environment.

In general, the method of the invention improves the underwater target detection effect by combining a plurality of factors such as adaptivity, feature extraction, k-means feature clustering and the like, and is particularly suitable for complex underwater scenes.

Drawings

In order to more clearly illustrate the invention or the technical solutions of the prior art, the following description will be given simply with reference to the accompanying drawings, which are used in the description of the embodiments or the prior art, it being evident that the following description is only one embodiment of the invention, and that other drawings can be obtained from these drawings without inventive effort for a person skilled in the art.

FIG. 1 is an overall logic block diagram of the underwater target classification detection method of the present invention.

Fig. 2 is a diagram of the BEMD frequency domain feature decomposition process in embodiment 1 of the present invention.

FIG. 3 is a schematic block diagram of a feature fusion mechanism of the present invention.

FIG. 4 is a graph showing the results of the k-means clustering algorithm in example 1 of the present invention.

FIG. 5 is a graph showing the results of the test of the target detecting portion in example 1 of the present invention.

Fig. 6 is a schematic diagram showing a simplified structure of the underwater target classification detection apparatus in embodiment 2.

Detailed Description

The invention will be further described with reference to specific examples.

Example 1:

in order to improve the accuracy and wide applicability of underwater low-light target detection, the invention provides an underwater target classification detection method based on two-dimensional empirical mode decomposition, and the whole process is shown in figure 1. The method adopts self-adaptive empirical mode decomposition (BEMD) to extract characteristics of input image data to be detected, and extracts characteristics of a hurst index, texture intensity and the like from a residual component sequence generated by BEMD decomposition. And then, establishing a fusion mechanism through multi-dimensional extraction features to form fusion feature data. And finally, establishing a k-means clustering algorithm model based on the fusion data, and carrying out classified target detection on the data to be detected by adopting the model, thereby obtaining a final detection result.

In this embodiment, an underwater fish target detection scenario is taken as an example, and the detection method of the present invention is further described.

1. Acquiring underwater fish detection image datasetAnd +.>The pretreatment, collection and pretreatment processes comprise:

(1) Collecting video data of an underwater object to be detected, and segmenting the collected image into pixelsShort video data +.2 s long>；

(2) At uniform time intervals, each segmentCut into image frame data of sequence length 50；

(3) For image frame data Corresponding label marking is carried out on the target to be detected, and the result is used as a group of data +.>；

(4) Repeatedly executing the above processes, and integrating the data processing results to obtain a data set。

2. Based on a two-dimensional empirical mode decomposition strategy and a model BEMD, the preprocessed data setAnd (3) performing intrinsic mode function decomposition (IMF) and extracting main frequency domain characteristic information from the image data information.

The process is shown in FIG. 2 and includes, for all image frame data in each set of sequence dataBEMD decomposition was performed (total of 50 images). And, the input of BEMD is defined as:

(1)

wherein,for the image sequence to be decomposed, < > for>And->Is the pixel coordinates in the image.

Thus, a set of dataAt the same time as input to obtain a sequenceAnd each coordinate pixel corresponds to the sequence +.>Wherein->. Calculating an envelope maximum value +.for each pixel coordinate for the corresponding sequence>And envelope minimum->Envelope average +.>The calculation of (2) is as shown in formula 2:

(2)

will beAs a result of the primary decomposition of BEMD +.>And the resulting residual component calculation is shown in equation 3:

(3)

judgingIf the decomposition termination condition is not satisfied, the residual component is continuously used as a new input signal, and the characteristic decomposition is repeatedly performed; on the contrary, if the termination condition is satisfied, the decomposition is stopped to finally obtain the total +.>The IMF component and the residual component (assuming co- +.>Sub-feature decomposition).

Wherein the decomposition termination conditions are specifically:

(4)

(5)。

3. residual component for BEMD decompositionExtracting features including Hurst index->Computation and texture intensity->Estimating; and integrating the multidimensional information through a feature fusion mechanism to obtain fusion feature data ++>。

Wherein, the Hurst indexThe acquisition process of (1) comprises:

the autocorrelation coefficient of the time-series data of the short-term correlation increases rapidly with time to decay to 0; whereas the autocorrelation coefficients of sequence data with long-term correlation decay more slowly. Thus, if the decay rate of the autocorrelation coefficients of a stationary sequence follows the power law decay (decays slower), then the time sequence has long memory. Based on this principle, the Hurst index (Hurst) is used to measure the degree of variation of the fluctuation range of the sequence data with time span, and indirectly reflects the long memory of the sequence data. Namely:

(6)

wherein,for the number of sequence data, here +.>；/>Is->The variation range of the observation points; />Is the standard deviation of these points. Thus->Indicate use +.>For->Normalization was performed. />Is constant, & lt>Is Hurst index, and +.>The method comprises the steps of carrying out a first treatment on the surface of the When->When there is no correlation in the sequence data; when->When the sequence data has long memory; />It indicates that the data has strong volatility.

Thus, the first and second substrates are bonded together,the calculation of (2) is expressed as:

(7)

Wherein, the texture strengthThe acquisition process and the characteristic data fusion mechanism of the system comprise:

(8)

wherein,representing a covariance taking operation.

(9)

where a and b are constant parameters that keep K positive. The fusion feature policy is defined as:

(10)

wherein,is +.>An IMF component->And the fusion characteristic data is corresponding. Thus for each group of initial data +.>Executing the feature extraction and fusion strategy to finally obtain a fusion frequency domain feature data set +.>. The data feature fusion process is shown in fig. 3.

4. Establishing a k-means clustering algorithm model, and adopting fusion characteristic dataAnd performing model training and testing to obtain the fish target classification detection model.

And taking the fusion characteristic data as an input characteristic vector of the object to be detected, and obtaining the total category number to be detected according to the category of the fish tag in the data set. Therefore, the technology presets the category number of the target to be detected asI.e. the number of preset clusters of the k-means algorithm is +.>. Performing k-means unsupervised learning training based on the fusion characteristic data set to finally obtain +.>Personal cluster center->The feature vector corresponding to the cluster center is a representative feature vector of such a target. The K-means cluster visualization result is shown in fig. 4, wherein different shape symbols represent fusion eigenvector data of different fishes.

5. Based on the underwater target classification detection model constructed by the construction method, carrying out target classification detection on the fish image to be detected, wherein the method comprises the following steps:

step 1, acquiring image sequence data of fish to be detected;

step 2, based on a two-dimensional empirical mode decomposition strategy and a model BEMD, performing intrinsic mode function decomposition IMF on the fish image data in the step 1, and extracting main frequency domain characteristic information from the image data information;

Step 4, fusing the characteristic dataInputting the characteristic classification detection model into the constructed underwater target classification detection model; the features areThe classification detection process comprises the following steps: calculate->Vector and->Euclidean distance of each cluster center is shown in formula 11:

(11)

(12)

and 5, outputting a fish target classification detection result.

Therefore, based on the target classification detection method, target classification detection is performed on the input fish sequence image (50 pieces of data), firstly, the input feature vector of the k-means algorithm is obtained after BEMD frequency domain feature extraction and feature fusion, and target classification and detection are performed based on the feature data, and the result is shown in fig. 5. The frame line is the visual rough marking result of the target detection, and the result verifies the practical feasibility of the method.

Example 2:

as shown in fig. 6, the present invention also provides a high-precision underwater object classification detection device, which comprises at least one processor and at least one memory, and also comprises a communication interface and an internal bus; the memory stores computer executing program; a computer-implemented program of the underwater target classification detection model constructed by the construction method described in embodiment 1 is stored in a memory; when the processor executes the computer executing program stored in the memory, the processor can be caused to execute a high-precision underwater target classification detection method. Wherein the internal bus may be an industry standard architecture (Industry Standard Architecture, ISA) bus, an external device interconnect (Peripheral Component, PCI) bus, or an extended industry standard architecture (. XtendedIndustry Standard Architecture, EISA) bus, among others. The buses may be divided into address buses, data buses, control buses, etc. For ease of illustration, the buses in the drawings of the present application are not limited to only one bus or one type of bus. The memory may include a high-speed RAM memory, and may further include a nonvolatile memory NVM, such as at least one magnetic disk memory, and may also be a U-disk, a removable hard disk, a read-only memory, a magnetic disk, or an optical disk.

The device may be provided as a terminal, server or other form of device.

Fig. 6 is a block diagram of an apparatus shown for illustration. The device may include one or more of the following components: a processing component, a memory, a power component, a multimedia component, an audio component, an input/output (I/O) interface, a sensor component, and a communication component. The processing component generally controls overall operation of the electronic device, such as operations associated with display, telephone calls, data communications, camera operations, and recording operations. The processing component may include one or more processors to execute instructions to perform all or part of the steps of the methods described above. Further, the processing component may include one or more modules that facilitate interactions between the processing component and other components. For example, the processing component may include a multimedia module to facilitate interaction between the multimedia component and the processing component.

The memory is configured to store various types of data to support operations at the electronic device. Examples of such data include instructions for any application or method operating on the electronic device, contact data, phonebook data, messages, pictures, videos, and the like. The memory may be implemented by any type of volatile or nonvolatile memory device or combination thereof, such as Static Random Access Memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disk.

The power supply assembly provides power to the various components of the electronic device. Power components may include a power management system, one or more power sources, and other components associated with generating, managing, and distributing power for electronic devices. The multimedia assembly includes a screen between the electronic device and the user that provides an output interface. In some embodiments, the screen may include a Liquid Crystal Display (LCD) and a Touch Panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from a user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensor may sense not only the boundary of a touch or slide action, but also the duration and pressure associated with the touch or slide operation. In some embodiments, the multimedia assembly includes a front camera and/or a rear camera. When the electronic device is in an operation mode, such as a shooting mode or a video mode, the front camera and/or the rear camera may receive external multimedia data. Each front camera and rear camera may be a fixed optical lens system or have focal length and optical zoom capabilities.

The audio component is configured to output and/or input an audio signal. For example, the audio component includes a Microphone (MIC) configured to receive external audio signals when the electronic device is in an operational mode, such as a call mode, a recording mode, and a voice recognition mode. The received audio signal may be further stored in a memory or transmitted via a communication component. In some embodiments, the audio assembly further comprises a speaker for outputting audio signals. The I/O interface provides an interface between the processing assembly and a peripheral interface module, which may be a keyboard, click wheel, button, etc. These buttons may include, but are not limited to: homepage button, volume button, start button, and lock button.

The sensor assembly includes one or more sensors for providing status assessment of various aspects of the electronic device. For example, the sensor assembly may detect an on/off state of the electronic device, a relative positioning of the assemblies, such as a display and keypad of the electronic device, a change in position of the electronic device or one of the assemblies of the electronic device, the presence or absence of user contact with the electronic device, an orientation or acceleration/deceleration of the electronic device, and a change in temperature of the electronic device. The sensor assembly may include a proximity sensor configured to detect the presence of nearby objects in the absence of any physical contact. The sensor assembly may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly may further include an acceleration sensor, a gyroscopic sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

The communication component is configured to facilitate communication between the electronic device and other devices in a wired or wireless manner. The electronic device may access a wireless network based on a communication standard, such as WiFi,2G, or 3G, or a combination thereof. In one exemplary embodiment, the communication component receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, the communication component further comprises a Near Field Communication (NFC) module to facilitate short range communications. For example, the NFC module may be implemented based on Radio Frequency Identification (RFID) technology, infrared data association (IrDA) technology, ultra Wideband (UWB) technology, bluetooth (BT) technology, and other technologies.

In an exemplary embodiment, the electronic device may be implemented by one or more Application Specific Integrated Circuits (ASICs), digital Signal Processors (DSPs), digital Signal Processing Devices (DSPDs), programmable Logic Devices (PLDs), field Programmable Gate Arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic elements for executing the methods described above.

Example 3:

the invention also provides a computer readable storage medium, wherein a computer executing program of the underwater target classification detection model constructed by the construction method according to the embodiment 1 is stored in the computer readable storage medium, and when the computer executing program is executed by a processor, a high-precision underwater target classification detection method can be realized.

In particular, a system, apparatus or device provided with a readable storage medium on which a software program code implementing the functions of any of the above embodiments is stored and whose computer or processor is caused to read and execute instructions stored in the readable storage medium may be provided. In this case, the program code itself read from the readable medium may implement the functions of any of the above-described embodiments, and thus the machine-readable code and the readable storage medium storing the machine-readable code form part of the present invention.

The storage medium may be implemented by any type or combination of volatile or nonvolatile memory devices such as Static Random Access Memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disks (e.g., CD-ROM, CD-R, CD-RW, DVD-20 ROM, DVD-RAM, DVD-RW), magnetic tape, and the like. A storage media may be any available media that can be accessed by a general purpose or special purpose computer.

It should be understood that the above processor may be a central processing unit (english: central Processing Unit, abbreviated as CPU), or may be other general purpose processors, digital signal processors (english: digital Signal Processor, abbreviated as DSP), application specific integrated circuits (english: application Specific Integrated Circuit, abbreviated as ASIC), or the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of a method disclosed in connection with the present invention may be embodied directly in a hardware processor for execution, or in a combination of hardware and software modules in a processor for execution.

It should be understood that a storage medium is coupled to the processor such the processor can read information from, and write information to, the storage medium. In the alternative, the storage medium may be integral to the processor. The processor and the storage medium may reside in an application specific integrated circuit (Application Specific Integrated Circuits, ASIC for short). The processor and the storage medium may reside as discrete components in a terminal or server.

The computer readable program instructions described herein may be downloaded from a computer readable storage medium to a respective computing/processing device or to an external computer or external storage device over a network, such as the internet, a local area network, a wide area network, and/or a wireless network. The network may include copper transmission cables, fiber optic transmissions, wireless transmissions, routers, firewalls, switches, gateway computers and/or edge servers. The network interface card or network interface in each computing/processing device receives computer readable program instructions from the network and forwards the computer readable program instructions for storage in a computer readable storage medium in the respective computing/processing device.

Computer program instructions for performing the operations of the present disclosure can be assembly instructions, instruction Set Architecture (ISA) instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source or object code written in any combination of one or more programming languages, including an object oriented programming language such as Smalltalk, c++ or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The computer readable program instructions may be executed entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computer (for example, through the Internet using an Internet service provider). In some embodiments, aspects of the present disclosure are implemented by personalizing electronic circuitry, such as programmable logic circuitry, field Programmable Gate Arrays (FPGAs), or Programmable Logic Arrays (PLAs), with state information of computer readable program instructions, which can execute the computer readable program instructions.

The foregoing description is only of the preferred embodiments of the present application and is not intended to limit the same, but rather, various modifications and variations may be made by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

While the foregoing describes the embodiments of the present invention, it should be understood that the present invention is not limited to the embodiments, and that various modifications and changes can be made by those skilled in the art without any inventive effort.

Claims

1. The method for building the high-precision underwater target classification detection model is characterized by comprising the following steps of:

2. The method for building a high-precision underwater target classification detection model according to claim 1, wherein the step S1 is specifically:

At uniform time intervals, each segmentCut into image frame data with a sequence length of 50 +.>；

3. The method for building a high-precision underwater target classification detection model according to claim 2, wherein the step S2 is specifically:

(1)

(2)

(3)

4. A method for building a high-precision underwater target classification detection model according to claim 3, wherein the decomposition termination condition is specifically:

(4)

(5)。

5. the method for building a high-precision underwater target classification detection model as claimed in claim 1, wherein the residual component for BEMD decomposition in S3Extracting features, wherein the Hurst index +.>The acquisition process comprises the following steps:

(6)

wherein,is the number of sequence data; />Is->The variation range of the observation points; />Is the standard deviation of these points, therefore +.>Indicate use +.>For->Performing standardization; />Is constant, & lt>Is Hurst index, and +.>The method comprises the steps of carrying out a first treatment on the surface of the When->When there is no correlation in the sequence data; when->When the sequence data has long memory; />It indicates that the data has strong volatility;

(7)

6. The method for building a high-precision underwater target classification detection model according to claim 1, wherein the texture intensity in S3 is as followsThe acquisition process and the characteristic data fusion mechanism of the system comprise:

(8)

wherein,representing a covariance extraction operation;

(9)

(10)

7. The method for building the high-precision underwater target classification detection model according to claim 1, wherein the training process of the k-means clustering algorithm model in S4 is as follows:

8. The high-precision underwater target classification detection method is characterized by comprising the following steps of:

step 1, acquiring image sequence data to be detected;

Step 4, fusing the characteristic dataInputting the characteristic classification detection model into an underwater target classification detection model constructed by the construction method according to any one of claims 1 to 7; the characteristic classification detection process comprises the following steps: calculate->Vector and->Euclidean distance of each cluster center is shown in formula 11:

(11)

(12)

and 5, outputting a target classification detection result.

9. A high-precision underwater target classification detection device is characterized in that: the apparatus includes at least one processor and at least one memory, the processor and the memory coupled; a computer-implemented program of an underwater target classification detection model constructed by the construction method according to any one of claims 1 to 7 is stored in the memory; when the processor executes the computer execution program stored in the memory, the processor is caused to execute a high-precision underwater target classification detection method.

10. A computer-readable storage medium, characterized by: the computer-readable storage medium stores therein a computer-executable program of the underwater target classification detection model constructed by the construction method according to any one of claims 1 to 7, which when executed by a processor, causes the processor to execute a high-precision underwater target classification detection method.