An underwater online intelligent monitoring and analyzing system and method

Abstract

This invention belongs to the field of underwater sea state data monitoring technology, and discloses an underwater online intelligent monitoring and analysis system and method. The method utilizes an Acoustic Doppler Current Profiler (ADCP) mounted on a seabed observation system to measure seabed currents, obtaining the velocity and direction of the water flow, and uses a binocular camera mounted on the same system to determine the size of underwater organisms. It employs an image recognition algorithm within the underwater observation system to perform image pattern recognition and classification of underwater organisms, identifying their species and quantity. Data is transmitted via an armored cable to the communication module of a virtual moored unmanned platform, and error control is implemented on the data transmitted via the armored cable to detect and correct errors that occur during transmission. The success of the returned information is determined through an information transmission control method within the underwater observation system. This invention enables image recognition of amphioxus in complex sea conditions, improving operational efficiency and safety.

Description

Technical Field

[0001] This invention belongs to the field of underwater sea state data monitoring technology, and particularly relates to an underwater online intelligent monitoring and analysis system and method. Background Technology

[0002] Traditional underwater image recognition systems cannot meet the requirements of complex sea conditions in terms of data transmission; underwater online intelligent monitoring and analysis systems consume too much energy and have intermittent power supply; traditional underwater image recognition systems can only identify images and cannot conduct comprehensive observation of the underwater environment.

[0003] Existing underwater monitoring and analysis systems can perform image analysis on seabed organisms, but this process lacks reliable data transmission, and the lack of a stable power supply often leads to signal interruptions and data loss during actual operation. Furthermore, the different carriers selected for existing underwater monitoring and analysis systems lack comprehensive observation capabilities.

[0004] Based on the above analysis, the problems and shortcomings of the existing technology are as follows:

[0005] Existing underwater online intelligent monitoring and analysis systems suffer from data transmission distortion, unsustainable power supply, and high energy consumption under complex sea conditions. Summary of the Invention

[0006] To overcome the problems existing in related technologies, the present invention discloses an underwater online intelligent monitoring and analysis system and method.

[0007] The technical solution is as follows: an underwater online intelligent monitoring and analysis method, comprising the following steps:

[0008] S1. The acoustic Doppler current profiler (ADCP) onboard the seabed observation system is used to measure the seabed current and obtain the speed and direction of the water flow. The binocular camera onboard the seabed observation system is used to determine the size of underwater organisms. The image recognition algorithm inside the underwater observation system is used to perform image pattern recognition and classification of underwater organisms to identify the species and quantity of underwater organisms.

[0009] S2, based on all acquired data and image information, uses armored cables to transmit data to the communication module of the virtual anchored unmanned platform, and performs error control on the data transmitted by the armored cables, detecting and correcting errors that occur during data transmission.

[0010] S3, in the process of information transmission back to the land base station by the communication module, the information transmission control method inside the underwater observation system is used to determine whether the back transmission is successful. If successful, the back transmission is performed; otherwise, steps S1-S2 are repeated until the back transmission is successful.

[0011] In step S1, the Acoustic Doppler Current Profiler (ADCP) performs seabed current measurement by emitting sound waves and receiving the reflected signals to measure the current velocity. As the sound waves propagate in the water, a Doppler frequency shift occurs due to the influence of the water flow, and the velocity and direction of the water flow are calculated. The Acoustic Doppler Current Profiler (ADCP) has multiple sound wave transmitters and uses beamforming and signal coherence processing algorithms to measure the current velocity at different depths, providing current velocity profile data.

[0012] Furthermore, the beamforming and signal coherence processing algorithm includes:

[0013] S101, Data Acquisition and Preprocessing;

[0014] Acoustic wave transmission and reception: ADCP emits multiple fixed-frequency acoustic beams into the water body, receives the reflected signals, and records the Doppler frequency shift;

[0015] Frequency shift extraction: Frequency shift is extracted using Fast Fourier Transform (FFT) or adaptive filtering algorithms.

[0016] Time-depth layering is used to layer the echo signal according to time windows, and combined with sound velocity profiles, the transmission frequency and beam tilt angle of each depth unit are collected.

[0017] S102, Radial velocity calculation;

[0018] The formula for calculating radial velocity is:

[0019]

[0020] In the formula, v r Δf is the radial velocity, Δf is the Doppler frequency shift, c is the speed of sound, f0 is the transmission frequency, and θ is the beam tilt angle.

[0021] S103, Three-dimensional flow velocity calculation;

[0022] The velocity components u, v, w are solved by an improved error-weighted least squares method.

[0023] The improved error-weighted least squares method takes into account the signal-to-noise ratio (SNR) differences of each beam, introduces a weight matrix W, and optimizes the flow velocity component calculation results. The expression is:

[0024]

[0025] In the formula, A is the beam pointing matrix, V r The radial velocity vector, * T It is the transpose matrix;

[0026] S104, Direction Calculation and Post-processing;

[0027] Flow velocity direction synthesis:

[0028] The eastward velocity u and the northward velocity v combine to form a horizontal flow velocity vector, with a direction angle of [missing information]. for:

[0029]

[0030] In the formula, This is the geomagnetic declination compensation value.

[0031] In step S1, the size of underwater organisms is determined using a binocular camera mounted on the seabed observation system, including:

[0032] (1) Calibrate the binocular camera and determine its internal and external parameters;

[0033] (2) By acquiring two images, left and right, feature extraction, feature matching, and disparity calculation are performed to achieve the perception of object depth information. An improved stereo matching algorithm is used to perceive object depth information, specifically including:

[0034] Step 1, Feature Extraction; Based on a deep learning-based feature extraction network, the outputs feature point coordinates and descriptors, expressed as:

[0035]

[0036] In the formula, F(I) represents the feature extraction image, and x i Let y be the x-coordinate of the feature point. i The ordinate of the feature point is... For feature point pixels, d i For high-dimensional descriptors, cross-perspective consistency is optimized through contrastive learning;

[0037] Step 2: Feature matching is performed using matching optimization based on graph neural networks;

[0038] Matching optimization based on graph neural networks includes: constructing a graph structure of feature points in the left and right images, where nodes are feature points and edges are descriptor similarity;

[0039] The matching score is optimized by aggregating neighborhood information through a graph attention network (GAT), as expressed in the following expression:

[0040]

[0041] In the formula, s ij To optimize the matching score, GAT() is an aggregation neighborhood function. Let i be the i-th high-dimensional descriptor in the left neighborhood. It is the j-th high-dimensional descriptor in the right neighborhood;

[0042] Differentiable soft matching (SoftMatch) is used to generate a probability matching matrix, which is then combined with epipolar constraints to filter interior points.

[0043] Step 3: Perform disparity calculation using improved cost aggregation and optimization;

[0044] Improved cost aggregation and optimization include:

[0045] Multi-scale cost volume construction: Multi-scale feature maps are generated using CNN to construct a 3D cost volume C(d,x,y);

[0046] Dynamic path-weighted smoothing (SGM) is employed, introducing edge-aware weights into the smoothing term of the SGM. The expression is as follows:

[0047]

[0048] In the formula, E(d) is the smoothing term of SGM, λ is the edge-aware weight, p and q are the p-th and q-th paths respectively, N is the number of paths, and w pq Let d be the edge sensing feature value of the p-th and q-th paths. p Let d be the high-dimensional descriptor of the p-th path. q w is the high-dimensional descriptor for the q-th path; pq =exp(-γ||I p -I q || 2 Based on the image gradient adaptive adjustment, γ is the focusing factor, exp() is the image gradient adaptive adjustment function, and I p Let I be the feature point of the p-th path. q These are the feature points of the q-th path;

[0049] Subpixel accuracy optimization, using parabolic interpolation or Newton's iteration method to optimize parallax:

[0050]

[0051] In the formula, d subpixel Let C be the sub-pixel precision optimization value, C be the multi-scale cost, and d be a certain high-dimensional descriptor;

[0052] (3) Reconstruct the three-dimensional scene based on the depth information.

[0053] In step S1, the underwater organisms are image pattern recognition and classification are performed using the image recognition algorithm within the underwater observation system, including:

[0054] First, the captured images are preprocessed with noise reduction and contrast enhancement.

[0055] Then, image features are extracted using a convolutional neural network (CNN).

[0056] Next, an attention mechanism and an optimized loss function are used to analyze the features and identify the species and quantity of underwater organisms. The attention mechanism adopts an improved Hybrid Attention Module (HAM), which includes channel attention, spatial attention, and feature fusion.

[0057] The channel attention includes: dynamically adjusting channel weights through global average pooling (GAP) and fully connected layers to enhance important feature channels, expressed as:

[0058]

[0059] W c =σ(W2·δ(W1·F) gap ))

[0060] In the formula, F gap W represents the feature map after global average pooling of the fully connected layer. c For the fully connected layer, H is the number of horizontal column types of the feature map, W is the number of vertical column types of the feature map of the fully connected layer, F(i,j) is the total feature map under the number of horizontal and vertical column types of the fully connected layer, F is the input feature map of the fully connected layer, σ is the Sigmoid function, δ is the ReLU activation function, and W1 and W2 are the parameters of the fully connected layer.

[0061] The spatial attention includes: generating a spatial weight map using convolutional layers to highlight biological regions, expressed as follows:

[0062] W s =σ(f 3×3 [F avg ,F max ])

[0063] In the formula, W s f represents the spatial weight value. 3×3 For a 3×3 convolution, F avg ,F max These are the results of channel averaging and max pooling, respectively.

[0064] The feature fusion includes: combining channels with spatial attention weights to enhance features, as expressed in the following expression:

[0065] F enhanced =W c ·W s ·F

[0066] In the formula, F enhanced is the feature enhancement value after combining channel and spatial attention weights, and F is the input feature map of the fully connected layer;

[0067] The optimized loss function adopts the improved composite loss function, Composite Loss, which includes classification task loss and quantity estimation loss.

[0068] The classification task loss employs an improved Focal Loss, which reduces the weight of easily classified samples to alleviate class imbalance. The expression is as follows:

[0069]

[0070] In the formula, L cls p is the loss value for the classification task. i To predict the probability, y i For real labels, a i Category weights;

[0071] The quantity estimation loss includes:

[0072] For densely populated biological regions, an adaptive Huber Loss is used to balance the robustness of outliers with the convergence rate. The expression is as follows:

[0073]

[0074] In the formula, L count To estimate the loss value for the quantity, n pred For the predicted value of dense biological outliers, n gt δ represents the actual value of the outlier points in the dense organisms, and δ is the ReLU activation function. δ is dynamically adjusted according to the training and decays with the number of iterations.

[0075] In step S2, the armored cable is used to transmit data to the communication module of the virtual moored unmanned platform. The underwater acoustic communication protocol includes the Modbus seabed communication protocol.

[0076] Error control is performed on the data transmitted through the armored cable to detect and correct errors that occur during data transmission. By adding a Cyclic Redundancy Check (CRC) code to the data, the receiver can detect whether errors have occurred during data transmission.

[0077] If an error is detected, the receiver requests the sender to resend the data;

[0078] When the receiver detects an error, it sends a request to the sender, asking the sender to retransmit the erroneous data packet. Upon receiving the request, the sender retransmits the packet until the receiver receives it correctly. In addition to sending the original data, the receiver also sends extra error correction codes when sending data. The receiver corrects the errors in the data based on these error correction codes.

[0079] In step S3, the information transmission control method within the underwater observation system includes:

[0080] Step 1: Determine the startup time. If yes, proceed to the next step; otherwise, re-determine the time.

[0081] Step 2: Start taking pictures and start the Acoustic Doppler Current Profiler (ADCP).

[0082] Step 3: Perform 4G backhaul;

[0083] Step 4: If the transmission is successful, put the device into sleep mode or return to step 1; if the transmission fails, determine whether to perform a retransmission timeout.

[0084] If the retransmission times out successfully, the system will either go into sleep mode or return to step 1. If the retransmission times out and fails, the system will return to step 3 and perform a 4G backhaul.

[0085] In step 4, the re-issuance timeout includes:

[0086] Step 4.1, Timeout threshold setting;

[0087] Dynamic timeout calculation, adjusting the timeout period based on historical network latency statistics:

[0088] T timeout =μRTT+3σRTT

[0089] In the formula, T timeout The timer period is μRTT, the average round-trip time is σRTT, and the standard deviation is σRTT, ensuring coverage of 95% of the fluctuation range;

[0090] Step 4.2, reissue trigger conditions; including:

[0091] If the initial transmission fails and no ACK is received during the 4G transmission in step 3, a retransmission process is triggered.

[0092] Retry queue management stores data to be retransmitted into a persistent queue and marks it as "to be retransmitted";

[0093] Step 4.3, reissue the execution process;

[0094] (1) Start the resend timer and set the timer period to T. timeout Start an asynchronous sending thread to attempt to resend data, while simultaneously listening for ACK responses;

[0095] (2) Wait for ACK or timeout, including: when ACK is successfully received, clear the corresponding data in the queue, terminate the timer, enter sleep or return to step 1; when ACK is not received after timeout, it is determined that the resend timeout has failed and the retry strategy is triggered.

[0096] Step 4.4, retry strategy optimization, includes: using the exponential backoff algorithm to successively increase the retry interval to avoid network congestion. The formula for the exponential backoff algorithm is:

[0097] T retry =min(T) base ×2 n-1 ,T max )

[0098] In the formula, T retry T is the optimal value for the retry strategy. base Based on the retry interval, T max T is the maximum retry interval. base =10s,T max =3600s, where n is the current number of retries.

[0099] Another object of the present invention is to provide an underwater online intelligent monitoring and analysis system, which implements the underwater online intelligent monitoring and analysis method, and the system includes:

[0100] The virtual mooring unmanned platform is used to stay on the sea surface and provide power to the seabed observation system and underwater observation system, and transmit the information processed by the seabed observation system and underwater observation system back to the land base station.

[0101] The seabed observation system transmits data with the virtual moored unmanned platform via armored cables, uses the onboard Acoustic Doppler Current Profiler (ADCP) to measure seabed currents, obtain the speed and direction of the water flow, and uses the onboard binocular camera to determine the size of underwater organisms.

[0102] The underwater observation system uses internal image recognition algorithms to perform image pattern recognition and classification of underwater organisms, identifying the species and quantity of underwater organisms; and in the communication module, it uses information transmission control methods to determine whether the backhauled information is successful when transmitting information back to the land base station.

[0103] The virtual anchoring unmanned platform consists of a buoy, a communication module, solar panels, a battery pack, a float, a counterweight, and a supply cable.

[0104] The buoy is responsible for carrying the equipment and keeping it on the sea surface;

[0105] The communication module is responsible for transmitting the collected data back to the land-based base station;

[0106] The power supply module includes solar panels and battery packs. When solar power is insufficient, the battery packs provide backup power.

[0107] The buoy, integrated with the buoy body, increases buoyancy;

[0108] Counterweights are used to increase stability;

[0109] The supply cable connects the buoy to the seabed observation system, transmitting data and power.

[0110] The seabed observation system includes: an acoustic Doppler current profiler (ADCP) for seabed current observation;

[0111] Four-color lights are used to provide visual signals at night or in low visibility conditions;

[0112] Binocular cameras, using stereo imaging technology, are used to observe the seabed environment;

[0113] Ultraviolet (UV) lights are used for specific purposes such as observing marine life or underwater navigation.

[0114] Batteries are used to provide electricity;

[0115] The data collection silo is used to store and process the collected data;

[0116] Seabed base, serving as a foundation, is fixed to the seabed;

[0117] Demand-side cable: Connects the submarine observation system to the supply-side cable, transmitting data and power;

[0118] The demand-side cable corresponds to the supply-side cable, running from the junction box to the seabed observation system, responsible for transmitting command data and power to the seabed observation system and receiving observation data;

[0119] The underwater observation system includes: an image recognition algorithm module, which analyzes and recognizes underwater images through software system operation;

[0120] The supporting algorithm software module that runs this image recognition algorithm is responsible for processing and analyzing image data to identify underwater targets.

[0121] Combining all the above technical solutions, the beneficial effects of this invention are as follows: This invention uses a virtual moored unmanned platform as the signal and energy supply end, and utilizes a docking unit to transmit data to the seabed observation system via armored cables for data transmission and energy supply. Simultaneously, the seabed observation system can be equipped with various sensors, particularly an ADCP sensor for current measurement, and a binocular camera to determine the size of underwater organisms. The underwater observation system uses machine vision algorithms for image pattern recognition and classification, and performs quantity counting of underwater organisms. Ultimately, it achieves the calculation of the number and size of underwater organisms.

[0122] This invention also solves the following problems:

[0123] Data transmission issues: Traditional underwater image recognition systems often experience data distortion in complex sea conditions, leading to signal interruptions and information loss. This problem is mainly due to the lack of a reliable data transmission mechanism and the instability of signal transmission in harsh marine environments. This invention utilizes a virtual moored unmanned platform as the signal and energy supply end, employing armored cables for data transmission and energy supply, thereby enhancing the stability and reliability of data transmission.

[0124] Energy consumption and supply issues: Existing systems consume excessive energy and suffer from discontinuous power supply. This is typically due to a lack of efficient energy management and supply systems, as well as the difficulty in providing continuous power in underwater environments. This invention provides a sustainable energy solution by integrating solar panels and battery packs to support the long-term operation of buoys.

[0125] Comprehensive Observation Capabilities: Traditional underwater image recognition systems can only identify images and cannot conduct comprehensive observations of the underwater environment. This limits the system's ability to comprehensively monitor changes in the marine environment. The underwater observation system of this invention not only includes image recognition algorithms but also utilizes a seabed observation system equipped with multiple sensors, such as an ADCP sensor for current measurement and a binocular camera to determine the size of underwater organisms, thereby achieving comprehensive monitoring and analysis of the underwater environment.

[0126] In summary, this invention enables image recognition of amphioxus under complex sea conditions, improving operational efficiency and safety. Furthermore, the method exhibits good adaptability and stability, making it suitable for various offshore operational scenarios. Attached Figure Description

[0127] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure;

[0128] Figure 1 This is a schematic diagram of the underwater online intelligent monitoring and analysis system provided in an embodiment of the present invention;

[0129] Figure 2 This is a schematic diagram of the stereo vision principle of a binocular camera provided in an embodiment of the present invention;

[0130] Figure 3 This is a schematic diagram of the image recognition algorithm provided in an embodiment of the present invention;

[0131] Figure 4 This is a schematic diagram of the information transmission control method inside the underwater observation system provided in this embodiment of the invention;

[0132] Figure 5 This is a schematic diagram of the error control principle provided in an embodiment of the present invention;

[0133] In the diagram: 1. Virtual moored unmanned platform; 2. Seabed observation system; 3. Underwater observation system. Detailed Implementation

[0134] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Many specific details are set forth in the following description to provide a thorough understanding of the present invention. However, the present invention can be practiced in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0135] Example 1, such as Figure 1 As shown, the underwater online intelligent monitoring and analysis system provided in this embodiment of the invention includes: a virtual mooring unmanned platform 1, a seabed observation system 2, and an underwater observation system 3.

[0136] The virtual moored unmanned platform 1 is used to maintain itself on the sea surface and to provide power to the seabed observation system 2 and the underwater observation system 3, as well as to transmit the information processed by the seabed observation system 2 and the underwater observation system 3 back to the land base station.

[0137] The seabed observation system 2 transmits data with the virtual moored unmanned platform 1 via armored cables.

[0138] The instrument uses the Acoustic Doppler Current Profiler (ADCP) to measure the seabed current, obtain the velocity and direction of the water flow, and uses the onboard binocular camera to determine the size of underwater organisms.

[0139] The underwater observation system 3 uses an internal image recognition algorithm to perform image pattern recognition and classification of underwater organisms, identifying the types and quantities of underwater organisms; and in the communication module, it uses information transmission control methods to determine whether the backhauled information is successful when transmitting information back to the land base station.

[0140] For example, the virtual mooring unmanned platform 1 consists of a buoy body, a communication module, a solar panel, a battery pack, a buoy, a counterweight, and a supply cable;

[0141] The seabed observation system 2 consists of a small snorkel, four-color lights, binocular cameras, ultraviolet lights, batteries, a data acquisition pod, a seabed base, and demand-side cables.

[0142] The underwater observation system 3 mainly consists of an image recognition algorithm module and a supporting algorithm software module that runs the image recognition algorithm.

[0143] In the virtual moored unmanned platform 1, the buoy body: as the main body of the virtual moored unmanned platform, the buoy body is responsible for carrying other equipment and keeping it on the sea surface.

[0144] Communication module: responsible for transmitting the collected data back to the land base station.

[0145] Power supply module (power supply system): includes solar panels and battery packs, in which the battery packs provide backup power when solar power is insufficient.

[0146] Buoy: Installed as an integral part of the buoy body, it increases buoyancy and ensures the stability of the platform on the sea surface.

[0147] Counterweights: Increase the stability of the virtual mooring unmanned platform and prevent it from being swept away by ocean currents.

[0148] Supply cable (towing cable): connects the buoy body to the seabed observation system 2, and serves to transmit data and power.

[0149] The buoy body, measuring 1.2 meters, serves as both the power supply module and the data transmission platform. The seabed observation system 2 is powered and controlled via the power supply module. The buoy body is equipped with three 100Ah battery packs for power supply.

[0150] The front end of the supply cable is connected to the power supply module and the communication module via a slip ring, and the rear end is connected to the data acquisition pod of the seabed observation system 2 via a tow cable connector; an anchor chain is also installed on the rear end of the supply cable.

[0151] For example, in the seabed observation system 2, the Acoustic Doppler Current Profiler (ADCP) (Xiao Kuolong) refers to a small seabed observation device.

[0152] Four-color lights: used to provide visual signals at night or in low visibility conditions.

[0153] Binocular camera: Used to observe the seabed environment through stereo imaging technology.

[0154] Ultraviolet light: May be used for specific marine life observation or underwater navigation.

[0155] Batteries: Provide power for the seabed observation system.

[0156] Data collection compartment (data connection box): Used to store and process collected data.

[0157] Seabed base (base platform): Serves as the base for the seabed observation system and is fixed to the seabed.

[0158] Demand-side cable: Connects the submarine observation system to the supply-side cable, transmitting data and power.

[0159] For example, an underwater binocular camera, an underwater light, and an ADCP are installed for monitoring. The underwater light and binocular camera are powered by the buoy, while the ADCP is self-powered. A data junction box is installed on the seabed base (base platform) for data aggregation and transmission.

[0160] In the data link, the link uses a supply-end cable (dragging cable) as the power supply and communication cable; a slip ring is installed at the top to prevent the cable from being twisted.

[0161] Regarding recycling and maintenance: additional lifting points will be added to the seabed base (base platform). Considering the potential for siltation during long-term deployment, divers can be arranged to install underwater lifting points and lift the platform if necessary.

[0162] Regarding power consumption: The buoy is powered by a 3*100Ah battery pack. Based on data transmission once a day (including image data and current measurement data collected by ADCP), the estimated operating time is 360 days without effective light-based charging.

[0163] For example, in the underwater observation system 3, the image recognition algorithm module analyzes and recognizes underwater images through the software system.

[0164] The supporting software module for running this image recognition algorithm is responsible for processing and analyzing image data to identify underwater targets.

[0165] Example 2: This embodiment of the invention provides an underwater online intelligent monitoring and analysis method, which includes:

[0166] S1. The acoustic Doppler current profiler (ADCP) onboard the seabed observation system is used to measure the seabed current and obtain the speed and direction of the water flow. The binocular camera onboard the seabed observation system is used to determine the size of underwater organisms. The image recognition algorithm inside the underwater observation system is used to perform image pattern recognition and classification of underwater organisms to identify the species and quantity of underwater organisms.

[0167] S2, based on all acquired data and image information, uses armored cables to transmit data to the communication module of the virtual anchored unmanned platform, and performs error control on the data transmitted by the armored cables, detecting and correcting errors that occur during data transmission.

[0168] S3, in the process of information transmission back to the land base station by the communication module, the information transmission control method inside the underwater observation system is used to determine whether the back transmission is successful. If successful, the back transmission is performed; otherwise, steps S1-S2 are repeated until the back transmission is successful.

[0169] For example, in step S1, the Acoustic Doppler Current Profiler (ADCP) measures the current velocity by emitting sound waves and receiving the reflected signals during seabed current measurement. When the sound waves propagate in the water, the Doppler frequency shift is generated by measuring the water flow, and the speed and direction of the water flow are calculated.

[0170] Specifically, an ADCP (Acoustic Doppler Current Profiler) is a commonly used ocean current measurement device. It measures current velocity by emitting sound waves and receiving the reflected signals. As sound waves propagate in water, they undergo a Doppler frequency shift due to the current; by measuring this shift, the ADCP can calculate the velocity and direction of the current. ADCPs typically have multiple sound wave transmitters, allowing simultaneous measurement of current velocities at different depths, thus providing current velocity profile data. To improve measurement accuracy, ADCPs may employ signal processing algorithms such as beamforming and coherent signal processing.

[0171] For example, in measuring flow velocity at different depths using beamforming and coherent signal processing algorithms, traditional methods calculate three-dimensional flow velocity through multi-beam radial velocity, while the improved beamforming and coherent signal processing algorithms of this invention combine signal processing optimization, error compensation, and coordinate system transformation correction to measure flow velocity at different depths. Specifically, this includes:

[0172] S101, Data Acquisition and Preprocessing;

[0173] Acoustic wave transmission and reception: ADCP emits multiple fixed-frequency acoustic beams (usually 4 beams, Janus configuration) into the water body, receives the reflected signals and records the Doppler frequency shift.

[0174] Frequency shift extraction utilizes Fast Fourier Transform (FFT) or adaptive filtering algorithms to extract frequency shifts and suppress noise (such as marine environmental noise or instrument electronic noise).

[0175] Time-depth layering is used to layer the echo signal according to time windows, and combined with the sound velocity profile (considering real-time changes in temperature, salinity and pressure), the transmission frequency and beam tilt angle of each depth unit are collected.

[0176] S102, Radial velocity calculation;

[0177] The formula for calculating radial velocity is:

[0178]

[0179] In the formula, v r Δf is the radial velocity, Δf is the Doppler frequency shift, c is the speed of sound, f0 is the transmission frequency, and θ is the beam tilt angle.

[0180] S103, Three-dimensional flow velocity calculation;

[0181] Traditional methods project the radial velocities of the four beams onto the instrument coordinate system (East-North-Sky, ENU), while this invention solves for the velocity components u, V, w using an improved error-weighted least squares method.

[0182] The improved error-weighted least squares calculation formula is as follows: Considering the signal-to-noise ratio (SNR) difference of each beam, a weight matrix W is introduced to optimize the flow velocity component calculation results:

[0183]

[0184] In the formula, A is the beam pointing matrix, V r The radial velocity vector, * T It is the transpose matrix;

[0185] S104, Direction Calculation and Post-processing;

[0186] Flow velocity direction synthesis:

[0187] The eastward velocity u and the northward velocity v combine to form a horizontal flow velocity vector, with a direction angle of [missing information]. for:

[0188]

[0189] In the formula, This is the geomagnetic declination compensation value (which needs to be updated in real time in conjunction with the geomagnetic field model).

[0190] In step S1, determining the size of underwater organisms using the binocular camera mounted on the seabed observation system includes:

[0191] (1) Calibrate the binocular camera and determine its internal and external parameters;

[0192] (2) By acquiring two images, left and right, feature extraction, feature matching and disparity calculation are performed to realize the perception of object depth information;

[0193] (3) Reconstruct the three-dimensional scene based on the depth information.

[0194] Specifically, the function of a binocular camera is to acquire the three-dimensional geometric information of an object based on the principle of stereo vision. The principle of binocular vision is based on the concept of parallax, which is the pixel displacement caused by the difference in viewpoints when two cameras observe the same object. By calculating parallax, the depth information of the object can be inferred. In binocular vision, the left camera is usually called the main camera, and the right camera is called the auxiliary camera. First, the binocular cameras need to be calibrated to determine their intrinsic and extrinsic parameters. Then, by acquiring the left and right images, feature extraction, feature matching, and parallax calculation are performed to perceive the depth information of the object. Finally, a three-dimensional scene is reconstructed based on the depth information. The principle diagram of binocular camera stereo vision is shown below. Figure 2 As shown.

[0195] For example, in step (2), the perception of object depth information is achieved by using an improved stereo matching algorithm, specifically including:

[0196] Step 1, Feature extraction (improved robust descriptors);

[0197] Deep learning-based feature extraction networks (such as SuperPoint) output feature point coordinates and descriptors:

[0198]

[0199] In the formula, F(I) represents the feature extraction image, and x i Let y be the x-coordinate of the feature point. i The ordinate of the feature point is... For feature point pixels, d i For high-dimensional descriptors, cross-perspective consistency is optimized through contrastive learning;

[0200] Step 2: Feature matching is performed using matching optimization based on graph neural networks (improved graph optimization matching);

[0201] The graph neural network-based matching optimization includes:

[0202] Construct a graph structure for feature points of the left and right images, where nodes are feature points and edges are descriptor similarities;

[0203] By aggregating neighborhood information through a graph attention network (GAT), the matching score is optimized.

[0204]

[0205] In the formula, s ij To optimize the matching score, GAT() is an aggregation neighborhood function. Let i be the i-th high-dimensional descriptor in the left neighborhood. It is the j-th high-dimensional descriptor in the right neighborhood;

[0206] Differentiable soft matching is used to generate a probability matching matrix, which is then combined with epipolar constraints to filter interior points.

[0207] Step 3: Perform disparity calculation using improved cost aggregation and optimization;

[0208] Improved cost aggregation and optimization include:

[0209] Multi-scale cost volume construction: Multi-scale feature maps are generated using CNN to construct a 3D cost volume C(d,x,y);

[0210] We employ dynamic path-weighted SGM, which includes introducing edge-aware weights into the smoothing term of traditional SGM:

[0211]

[0212] In the formula, E(d) is the smoothing term of SGM, λ is the edge-aware weight, p and q are the p-th and q-th paths respectively, N is the number of paths, and w pq Let d be the edge sensing feature value of the p-th and q-th paths. p Let d be the high-dimensional descriptor of the p-th path. q w is the high-dimensional descriptor for the q-th path; pq =exp(-γ||I p -I q || 2 Based on the image gradient adaptive adjustment, γ is the focusing factor, exp() is the image gradient adaptive adjustment function, and I p Let I be the feature point of the p-th path. q These are the feature points of the q-th path;

[0213] Subpixel accuracy optimization, using parabolic interpolation or Newton's iteration method to optimize parallax:

[0214]

[0215] In the formula, d subpixel Let C be the sub-pixel precision optimization value, C be the multi-scale cost, and d be a certain high-dimensional descriptor;

[0216] For example, in step S1, the image pattern recognition and classification of underwater organisms using the image recognition algorithm within the underwater observation system includes:

[0217] First, the captured image is preprocessed with noise reduction and contrast enhancement.

[0218] Then, image features are extracted using a convolutional neural network (CNN).

[0219] Next, an attention mechanism and an optimized loss function are used to analyze the features and identify the types and quantities of underwater organisms.

[0220] Specifically, underwater biometrics algorithms are typically based on machine vision technology, using image recognition algorithms to perform pattern recognition and classification of underwater organisms. These algorithms may employ deep learning models, such as the YOLO (You Only LookOnce) series, which enables real-time object detection and recognition. To adapt to the unique characteristics of the underwater environment, such as changing light and water scattering, image recognition algorithms incorporate attention mechanisms and optimized loss functions. During the recognition process, the captured images are first preprocessed, such as denoising and contrast enhancement, and then image features are extracted using a convolutional neural network (CNN). Next, a classifier is used to analyze the features and identify the species and quantity of underwater organisms. For size measurement, image recognition may utilize stereo images acquired by binocular cameras, calculating the three-dimensional dimensions of the organisms using stereo vision technology. A schematic diagram of the image recognition algorithm is shown below. Figure 3 As shown.

[0221] For example, in the process of using a classifier to introduce an attention mechanism and optimize the loss function to analyze features and identify the types and quantities of underwater organisms, the attention mechanism adopts an improved Hybrid Attention Module (HAM), which includes channel attention, spatial attention, and feature fusion.

[0222] The channel attention includes:

[0223] Enhance important feature channels by dynamically adjusting channel weights through global average pooling (GAP) and fully connected layers:

[0224]

[0225] W c =σ(W2·δ(W1·F) gap ))

[0226] In the formula, F gap W represents the feature map after global average pooling of the fully connected layer. c For the fully connected layer, H is the number of horizontal column types of the feature map, W is the number of vertical column types of the feature map of the fully connected layer, F(i,j) is the total feature map under the number of horizontal and vertical column types of the fully connected layer, F is the input feature map of the fully connected layer, σ is the Sigmoid function, δ is the ReLU activation function, and W1 and W2 are the parameters of the fully connected layer.

[0227] Spatial attention includes:

[0228] Spatial weight maps are generated using convolutional layers to highlight biological regions. The expression is as follows:

[0229] W s =σ(f 3×3 [F avg ,F max ])

[0230] In the formula, W s f represents the spatial weight value. 3×3 For a 3×3 convolution, F avg ,F max These are the results of channel averaging and max pooling, respectively.

[0231] The feature fusion includes: combining channels with spatial attention weights to enhance features, as expressed in the following expression:

[0232] F enhanced =W c ·W s ·F

[0233] In the formula, F enhanced is the feature enhancement value after combining channel and spatial attention weights, and F is the input feature map of the fully connected layer;

[0234] The optimized loss function adopts an improved composite loss function, which includes classification task loss and quantity estimation loss.

[0235] The classification task loss employs an improved Focal Loss, which reduces the weight of easily classified samples to alleviate class imbalance. The expression is as follows:

[0236]

[0237] In the formula, L cls p is the loss value for the classification task. i To predict the probability, y i For real labels, a i The class weights are inversely proportional to the number of samples.

[0238] The quantity estimation loss includes:

[0239] For densely populated biological regions, an adaptive Huber Loss is used to balance the robustness of outliers with the convergence rate. The expression is as follows:

[0240]

[0241] In the formula, L count To estimate the loss value for the quantity, n pred For the predicted value of dense biological outliers, n gtδ represents the actual value of the outlier points in the dense organisms, and δ is the ReLU activation function. δ is dynamically adjusted according to the training and decays with the number of iterations.

[0242] For example, in step S2, when the armored cable is used to transmit data to the communication module of the virtual moored unmanned platform, the underwater acoustic communication protocol includes the Modbus seabed communication protocol.

[0243] Error control is performed on the data transmitted through the armored cable. The detection and correction of errors that occur during data transmission include: adding a Cyclic Redundancy Check (CRC) code to the data so that the receiver can detect whether errors have occurred during data transmission.

[0244] If an error is detected, the receiver requests the sender to resend the data;

[0245] When the receiver detects an error, it sends a request to the sender, asking the sender to retransmit the erroneous data packet. Upon receiving the request, the sender retransmits the packet until the receiver receives it correctly. In addition to sending the original data, the receiver also sends extra error correction codes when sending data. The receiver corrects the errors in the data based on these error correction codes.

[0246] Specifically, in terms of data transmission technology, the virtual moored unmanned platform 1 transmits data to the seabed observation system 2 via an armored cable. The armored cable is a robust cable typically used in marine environments, capable of withstanding seawater corrosion and seabed pressure.

[0247] Because electromagnetic waves propagate much less efficiently underwater than sound waves, underwater acoustic communication has become the primary method of underwater communication. Underwater acoustic communication protocols typically include modulation and demodulation techniques, signal processing algorithms, and data link layer protocols to ensure reliable data transmission in complex underwater environments.

[0248] Data transmission employs multiple communication protocols, such as the Modbus dedicated submarine communication protocol, to ensure stable and reliable data transmission. Simultaneously, error control and flow control technologies may be used to improve transmission efficiency and reduce data loss.

[0249] Error control is a mechanism used to detect and correct errors that occur during data transmission. Its purpose is to ensure that the data received by the receiver is completely consistent with the data sent by the sender, even if errors occur during transmission. By adding a Cyclic Redundancy Check (CRC) code to the data, the receiver can detect whether errors have occurred during transmission. If an error is detected, the receiver can request the sender to retransmit the data. When an error is detected, the receiver sends a request to the sender, requesting the sender to retransmit the erroneous data packet. Upon receiving the request, the sender will retransmit the packet until the receiver receives it correctly. In addition to sending the original data, additional error correction codes are also sent during transmission. The receiver can use these error correction codes to correct errors in the data, such as forward error correction (FEC). Figure 5 Error control principle diagram.

[0250] For example, in step S3, such as Figure 4 As shown, the information transmission control methods within the underwater observation system include:

[0251] Step 1: Determine the startup time. If yes, proceed to the next step; otherwise, re-determine the time.

[0252] Step 2: Start taking pictures and start the Acoustic Doppler Current Profiler (ADCP).

[0253] Step 3: Perform 4G backhaul;

[0254] Step 4: If the backhaul is successful, put the device into sleep mode or return to step 1; if the backhaul fails, determine whether to perform a retransmission timeout. If the retransmission timeout is successful, put the device into sleep mode or return to step 1; if the retransmission timeout fails, return to step 3 and perform a 4G backhaul.

[0255] For example, the resend timeout includes:

[0256] Step 4.1, Timeout threshold setting;

[0257] Dynamic timeout calculation, adjusting the timeout period based on historical network latency statistics (such as sliding window average RTT):

[0258] T timeout =μRTT+3σRTT

[0259] In the formula, T timeout The timer period is μRTT, the average round-trip time is σRTT, and the standard deviation is σRTT, ensuring coverage of 95% of the fluctuation range;

[0260] Step 4.2, reissue trigger conditions; including:

[0261] If the first transmission fails and no ACK is received during the 4G transmission in step 3 (or the HTTP status code is not 200), the retransmission process is triggered.

[0262] Retry queue management stores data to be retransmitted in a persistent queue (such as Flash storage) and marks it as "to be retransmitted".

[0263] Step 4.3, reissue the execution process;

[0264] 1) Start the resend timer and set the timer period to T. timeout ;

[0265] Start an asynchronous sending thread to attempt to resend data, while simultaneously listening for ACK responses;

[0266] 2) Waiting for ACK or timeout, including:

[0267] Upon successful receipt of an ACK, clear the corresponding data in the queue, terminate the timer, and enter sleep mode or return to step 1.

[0268] If no ACK is received within the timeout period, it is determined that the retransmission timeout has failed and the retry policy is triggered.

[0269] Step 4.4, retry strategy optimization, including:

[0270] The exponential backoff algorithm, by progressively increasing the retry interval, avoids network congestion. The formula for calculating the exponential backoff algorithm is as follows:

[0271] T retry =min(T) base ×2 n-1 ,T max )

[0272] In the formula, T retry T is the optimal value for the retry strategy. base Based on the retry interval, T max T is the maximum retry interval. base =10s,T max =3600s, where n is the current number of retries.

[0273] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0274] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by those skilled in the art within the scope of the technology disclosed in the present invention and within the spirit and principles of the present invention should be covered within the scope of protection of the present invention.

Claims

1. An online intelligent monitoring analysis method under water, characterized in that, The method includes the following steps: S1. The acoustic Doppler current profiler (ADCP) onboard the seabed observation system is used to measure the seabed current and obtain the speed and direction of the water flow. The binocular camera onboard the seabed observation system is used to determine the size of underwater organisms. The image recognition algorithm inside the underwater observation system is used to perform image pattern recognition and classification of underwater organisms to identify the species and quantity of underwater organisms. S2, based on all acquired data and image information, uses armored cables to transmit data to the communication module of the virtual anchored unmanned platform, and performs error control on the data transmitted by the armored cables, detecting and correcting errors that occur during data transmission. S3, In the process of information transmission back to the land base station by the communication module, the information transmission control method inside the underwater observation system is used to determine whether the back transmission is successful. If successful, the back transmission is performed; otherwise, steps S1-S2 are repeated until the back transmission is successful. In step S1, the underwater organisms are image pattern recognition and classification are performed using the image recognition algorithm within the underwater observation system, including: First, the captured images are preprocessed with noise reduction and contrast enhancement. Then, image features are extracted using a convolutional neural network (CNN). Next, an attention mechanism and an optimized loss function are used to analyze the features and identify the species and quantity of underwater organisms. The attention mechanism adopts an improved Hybrid Attention Module (HAM), which includes channel attention, spatial attention, and feature fusion. The channel attention includes: dynamically adjusting channel weights through global average pooling (GAP) and fully connected layers to enhance important feature channels, expressed as: ； ； In the formula, is the feature channel of the full connection layer after global average pooling, is the enhanced feature channel of the full connection layer after multi-scale cost processing, is the number of horizontal categories of the feature map, is the number of vertical categories of the feature map of the full connection layer, is the total feature map of the full connection layer horizontal category number and vertical category number, is the input feature map of the full connection layer, is the Sigmoid function, is the ReLU activation function, is the full connection layer parameter; The spatial attention includes: generating a spatial weight map using convolutional layers to highlight biological regions, expressed as follows: ； wherein, is a spatial weight value, is a 3x3 convolution, are the channel average and max pooling results, respectively; The feature fusion includes: combining channels with spatial attention weights to enhance features, as expressed in the following expression: ； In the formula, is the feature enhancement value after the channel and spatial attention weights are combined, is the full connection layer input feature map; The optimized loss function adopts the improved composite loss function, Composite Loss, which includes classification task loss and quantity estimation loss. The classification task loss employs an improved Focal Loss, which reduces the weight of easily classified samples to alleviate class imbalance. The expression is as follows: ； In the formula, is a classification task loss value, is a prediction probability, is a true label, is a class weight; The quantity estimation loss includes: For densely populated biological regions, an adaptive Huber Loss is used to balance the robustness of outliers with the convergence rate. The expression is as follows: ； wherein, is the quantity estimated loss value, is the dense biological outlier predicted value, is the dense biological outlier actual value, is the ReLU activation function, is adjusted dynamically according to the training, and decays with the number of iterations; In step S3, the information transmission control method within the underwater observation system includes: Step 1: Determine the startup time. If yes, proceed to the next step; otherwise, re-determine the time. Step 2: Start taking pictures and start the Acoustic Doppler Current Profiler (ADCP). Step 3: Perform 4G backhaul; Step 4: If the transmission is successful, put the device into sleep mode or return to step 1; if the transmission fails, determine whether to perform a retransmission timeout. If the retransmission times out successfully, the system will either go into sleep mode or return to step 1. If the retransmission times out and fails, the system will return to step 3 and perform a 4G backhaul.

2. The method of claim 1, wherein, In step S1, the Acoustic Doppler Current Profiler (ADCP) performs seabed current measurement by emitting sound waves and receiving the reflected signals to measure the current velocity. As the sound waves propagate in the water, a Doppler frequency shift occurs due to the influence of the water flow, and the velocity and direction of the water flow are calculated. The Acoustic Doppler Current Profiler (ADCP) has multiple sound wave transmitters and uses beamforming and signal coherence processing algorithms to measure the current velocity at different depths, providing current velocity profile data.

3. The underwater online intelligent monitoring and analysis method according to claim 2, characterized in that, The beamforming and signal coherence processing algorithm includes: S101, Data Acquisition and Preprocessing; Acoustic wave transmission and reception: ADCP emits multiple fixed-frequency acoustic beams into the water body, receives the reflected signals, and records the Doppler frequency shift; Frequency shift extraction: Frequency shift is extracted using Fast Fourier Transform (FFT) or adaptive filtering algorithms. Time-depth layering is used to layer the echo signal according to time windows, and combined with sound velocity profiles, the transmission frequency and beam tilt angle of each depth unit are collected. S102, Radial velocity calculation; The formula for calculating radial velocity is: ； In the formula, Radial velocity, For Doppler frequency shift, For the speed of sound, For the transmission frequency, The beam tilt angle; S103, Three-dimensional flow velocity calculation; The velocity components are solved using an improved error-weighted least squares method. ; The improved error-weighted least squares method takes into account the signal-to-noise ratio (SNR) differences of each beam and introduces a weight matrix. The optimized solution for the velocity component is expressed as follows: ； In the formula, For the beam pointing matrix, The radial velocity vector, It is the transpose matrix; S104, Direction Calculation and Post-processing; Flow velocity direction synthesis: Eastward speed and northbound speed Composite horizontal velocity vector, direction angle for: ； In the formula, This is the geomagnetic declination compensation value.

4. The underwater online intelligent monitoring and analysis method according to claim 1, characterized in that, In step S1, the size of underwater organisms is determined using a binocular camera mounted on the seabed observation system, including: (1) Calibrate the binocular camera and determine its internal and external parameters; (2) By acquiring the left and right images, feature extraction, feature matching, and disparity calculation are performed to realize the perception of object depth information. An improved stereo matching algorithm is used to perceive object depth information, specifically including: Step 1, Feature Extraction; Based on a deep learning-based feature extraction network, the outputs feature point coordinates and descriptors, expressed as: ； In the formula, For feature extraction images, The x-coordinate of the feature point The ordinate of the feature point is... For feature point pixels, For high-dimensional descriptors, cross-perspective consistency is optimized through contrastive learning; Step 2: Feature matching is performed using matching optimization based on graph neural networks; Matching optimization based on graph neural networks includes: constructing a graph structure of feature points in the left and right images, where nodes are feature points and edges are descriptor similarity; The matching score is optimized by aggregating neighborhood information through a graph attention network (GAT), as expressed in the following expression: ； In the formula, To optimize the matching score, For aggregation neighborhood function, For the left neighbor A high-dimensional descriptor, For the right neighbor number A high-dimensional descriptor; Differentiable soft matching (SoftMatch) is used to generate a probability matching matrix, which is then combined with epipolar constraints to filter interior points. Step 3: Perform disparity calculation using improved cost aggregation and optimization; Improved cost aggregation and optimization include: Multi-scale cost volume construction: Multi-scale feature maps are generated using CNN to construct a 3D cost volume. ; Dynamic path-weighted smoothing (SGM) is employed, introducing edge-aware weights into the smoothing term of the SGM. The expression is as follows: ； In the formula, For the smoothing term of SGM, For edge-aware weights, The first One path, For the number of paths, For the first Edge-aware feature values ​​of each path For the first A high-dimensional descriptor for each path For the first A high-dimensional descriptor for each path; Adaptive adjustment based on image gradient As a focusing factor, This is an image gradient adaptive adjustment function. For the first Feature points of each path For the first Feature points of each path; Subpixel accuracy optimization, using parabolic interpolation or Newton's iteration method to optimize parallax: ； In the formula, This is a sub-pixel precision optimized value. For multi-scale costs, For a certain high-dimensional descriptor; (3) Reconstruct the three-dimensional scene based on the depth information.

5. The underwater online intelligent monitoring and analysis method according to claim 1, characterized in that, In step S2, the armored cable is used to transmit data to the communication module of the virtual moored unmanned platform. The underwater acoustic communication protocol includes the Modbus seabed communication protocol. Error control is performed on the data transmitted through the armored cable to detect and correct errors that occur during data transmission. By adding a Cyclic Redundancy Check (CRC) code to the data, the receiver can detect whether errors have occurred during data transmission. If an error is detected, the receiver requests the sender to resend the data; When the receiver detects an error, it sends a request to the sender, asking the sender to retransmit the erroneous data packet. Upon receiving the request, the sender retransmits the packet until the receiver receives it correctly. In addition to sending the original data, the receiver also sends extra error correction codes when sending data. The receiver corrects the errors in the data based on these error correction codes.

6. The underwater online intelligent monitoring and analysis method according to claim 1, characterized in that, In step 4, the re-issuance timeout includes: Step 4.1, Timeout threshold setting; Dynamic timeout calculation, adjusting the timeout period based on historical network latency statistics: ； In the formula, For timer period, This represents the average round-trip time. The standard deviation is used to ensure coverage of 95% of the fluctuation range; Step 4.2, reissue trigger conditions; including: If the initial transmission fails and no ACK is received during the 4G transmission in step 3, a retransmission process is triggered. Retry queue management stores data to be retransmitted into a persistent queue and marks it as "to be retransmitted"; Step 4.3, reissue the execution process; (1) Start the resend timer and set the timer period to . Start an asynchronous sending thread to attempt to resend data, while simultaneously listening for ACK responses; (2) Waiting for ACK or timeout, including: when ACK is successfully received, clear the corresponding data in the queue, terminate the timer, enter sleep or return to step 1; when ACK is not received after timeout, it is determined that the resend timeout has failed and the retry strategy is triggered. Step 4.4, retry strategy optimization, includes: using the exponential backoff algorithm to successively increase the retry interval to avoid network congestion. The formula for the exponential backoff algorithm is: ； In the formula, The optimized value for the retry strategy, Based on the retry interval, This is the maximum retry interval. , This represents the current number of retries.

7. An underwater online intelligent monitoring and analysis system, characterized in that, The system implements the underwater online intelligent monitoring and analysis method as described in any one of claims 1-6, and the system includes: The virtual mooring unmanned platform (1) is used to stay on the sea surface and provide power to the seabed observation system (2) and the underwater observation system (3), and transmit the information processed by the seabed observation system (2) and the underwater observation system (3) back to the land base station; The seabed observation system (2) transmits data with the virtual moored unmanned platform (1) through armored cables, uses the onboard acoustic Doppler current profiler ADCP to measure the seabed current, obtain the speed and direction of the water flow, and uses the onboard binocular camera to determine the size of underwater organisms. The underwater observation system (3) uses the image recognition algorithm inside the underwater observation system to perform image pattern recognition and classification of underwater organisms, and identifies the types and quantities of underwater organisms; and in the communication module, it uses the information transmission control method to determine whether the information transmission is successful in transmitting information back to the land base station.

8. The underwater online intelligent monitoring and analysis system according to claim 7, characterized in that, The virtual anchored unmanned platform (1) consists of a buoy body, a communication module, solar panels, a battery pack, a float, a counterweight, and a supply cable; The buoy is responsible for carrying the equipment and keeping it on the sea surface; The communication module is responsible for transmitting the collected data back to the land-based base station; The power supply module includes solar panels and battery packs. When solar power is insufficient, the battery packs provide backup power. The buoy, integrated with the buoy body, increases buoyancy; Counterweights are used to increase stability; The supply cable connects the buoy body to the seabed observation system (2) to transmit data and power; The seabed observation system (2) includes: an acoustic Doppler current profiler (ADCP) for seabed current observation; Four-color lights are used to provide visual signals at night or in low visibility conditions; Binocular cameras, using stereo imaging technology, are used to observe the seabed environment; Ultraviolet (UV) lights are used for specific purposes such as observing marine life or underwater navigation. Batteries are used to provide electricity; The data collection silo is used to store and process the collected data; Seabed base, serving as a foundation, is fixed to the seabed; Demand-side cable: Connects the submarine observation system to the supply-side cable, transmitting data and power; The demand-side cable corresponds to the supply-side cable, running from the junction box to the seabed observation system, responsible for transmitting command data and power to the seabed observation system and receiving observation data; The underwater observation system (3) includes: an image recognition algorithm module, which analyzes and recognizes underwater images through a software system; The supporting algorithm software module that runs this image recognition algorithm is responsible for processing and analyzing image data to identify underwater targets.

Images