Mariculture net cage management and control system

Abstract

The invention provides a comprehensive and intelligent mariculture net cage management and control system. The system integrates a multi-parameter sensor module, an underwater robot module, an underwater camera module, a central control module, a data storage and analysis module, an early warning module, a remote communication module and an automatic execution module. By monitoring multiple environmental parameters such as water temperature, salinity, dissolved oxygen and the like in real time and utilizing the underwater robot and the underwater camera to carry out autonomous inspection and fish behavior observation, the system can accurately master the culture environment condition. The AI algorithm and the intelligent model are combined, the system dynamically adjusts the bait casting amount, the oxygenation frequency and the net cage position, and the breeding benefits are effectively improved. Meanwhile, the early warning module carries out graded early warning on risk events such as typhoon and diseases, and the breeding safety is ensured. The remote communication and visual human-computer interface further enhances the remote monitoring and management capability of the system.

Description

Technical Field

[0001] The present invention relates to the technical field of aquaculture cage control systems, and in particular to a seawater aquaculture cage control system. Background Art

[0002] With the rapid development of the seawater aquaculture industry, traditional aquaculture methods have been difficult to meet the modern and large-scale aquaculture needs. Currently, seawater aquaculture faces many challenges such as inaccurate monitoring of environmental parameters, low efficiency of cage inspection, and untimely disease prevention and control. The marine environment is complex and changeable, and manual monitoring is difficult to cover comprehensively and is easily affected by subjective factors, resulting in incorrect aquaculture decisions. At the same time, traditional cage inspections rely on manual diving or vessel operations, which are costly, inefficient, and pose safety hazards. In addition, the outbreak of diseases is often sudden, and traditional prevention and control measures are difficult to achieve timely early warning and effective treatment, causing huge economic losses to farmers. Therefore, developing an intelligent seawater aquaculture cage control system to achieve comprehensive monitoring of the aquaculture environment, precise management, and effective prevention and control of risks has become an urgent need to enhance the competitiveness of the seawater aquaculture industry. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to provide a seawater aquaculture cage control system to solve at least the above problems.

[0004] The technical solution adopted by the present invention is as follows: A seawater aquaculture cage control system, comprising: A multi-parameter sensor module, arranged inside and around the aquaculture cage, for real-time monitoring of water temperature, salinity, dissolved oxygen, pH value, ammonia nitrogen concentration, turbidity, and water flow velocity, and transmitting the collected data to the central control module through wireless communication; An underwater robot module, equipped with a high-definition camera, a robotic arm, and an environmental sensor, for autonomously inspecting the integrity of the cage structure, cleaning the attachments on the netting, troubleshooting, and observing fish behavior. The movement path of the robot is dynamically optimized by the central control module based on real-time environmental data; An underwater camera module, using anti-corrosion high-definition imaging equipment, deployed in key areas of the aquaculture cage, for real-time shooting of fish activity status, feeding conditions, and body surface health status. The image data is analyzed by an AI algorithm and then fed back to the central control module; A central control module, integrated with an edge computing unit, for receiving and processing multi-source monitoring data, generating an optimized control strategy through built-in environmental dynamic models and fish growth models, and dynamically adjusting the feeding amount, oxygenation frequency, and cage position; A data storage and analysis module, constructing a distributed database to store historical environmental data, fish growth records, and equipment operation logs, and mining the correlation rules between environmental parameters and aquaculture benefits in combination with machine learning algorithms to support long-term trend prediction; The early warning module, based on threshold determination and pattern recognition technologies, classifies and gives early warnings for typhoon, red tide, hypoxia, and disease risk events, and pushes them to the management personnel through multi-channels of sound and light alarms, text messages, and platform pop-ups; The remote communication module uses 4G / 5G and satellite communication dual links to remotely and real-time transmit monitoring data, control instructions, and early warning information, and supports collaborative access among shore-based monitoring centers, mobile terminals, and cloud platforms; The automatic execution module includes a precise feeding machine driven by a servo motor, a variable frequency aeration pump, a cage mooring adjustment device, and a net cleaning mechanism, and performs closed-loop control operations according to the instructions of the central control module.

[0005] Furthermore, the multi-parameter sensor module adopts a modular design. The sensor nodes can be quickly replaced through waterproof connectors, and are equipped with a self-cleaning mechanism to prevent biological attachment from affecting the measurement accuracy; The sensor data fusion spatio-temporal interpolation algorithm reconstructs the missing values through the data of surrounding nodes in case of single-point failure to ensure the continuity of monitoring.

[0006] Furthermore, the underwater robot module adopts a bionic propulsion system, equipped with a multi-degree-of-freedom robotic arm and a laser calibration device, which can autonomously identify the damaged position of the net and perform repair operations; The side-scan sonar and optical camera carried by the robot work together to build a three-dimensional digital twin model of the cage for the decision-making reference of the central control module.

[0007] Furthermore, the central control module is built-in with a multi-objective optimization algorithm, which takes into account factors such as the fish growth stage, residual bait amount, water temperature, and sea current speed in the bait feeding decision-making, and dynamically generates a feeding plan; The module integrates a reinforcement learning framework, and continuously optimizes the control strategy through historical operation data and feedback on aquaculture benefits.

[0008] Furthermore, in the typhoon early warning subsystem of the early warning module, meteorological forecast data and the cage structure dynamics model are integrated to calculate the force distribution of the cage under different wind and wave levels, and trigger reinforcement or disaster avoidance plans in advance; Disease early warning analyzes the fish body surface image features through a convolutional neural network, identifies common diseases such as white spot disease and fin rot disease, and marks suspected infected individuals.

[0009] Furthermore, the remote communication module uses the LoRaWAN protocol to form an underwater sensor self-organizing network, and the relay buoy is equipped with a Beidou short message terminal to achieve offshore communication guarantee; The data transmission process is encrypted using the national secret SM4 algorithm, and the communication link has a heartbeat detection and automatic switching function.

[0010] Furthermore, the feeder of the automated execution module is equipped with a weighing sensor and an image recognition unit to correct the feeding amount in real time and count the feeding activity. The net cage position adjustment device controls the retraction and release of the anchor chain through an electric winch, and actively avoids bad water masses in combination with ocean current forecast data.

[0011] Furthermore, it includes a visual human-machine interface, which displays the distribution of net cages, the heat map of environmental parameters, and the operating status of equipment on a 3D GIS map. The visual human-machine interface integrates a virtual reality (VR) operation mode, supporting remote first-person perspective inspection and manual intervention of key equipment.

[0012] Compared with the prior art, the beneficial effects of the present invention are as follows: The seawater aquaculture net cage control system provided by the present invention realizes the refined management of the seawater aquaculture environment by integrating multi-source monitoring and intelligent control technologies. The system can monitor and accurately analyze environmental parameters in real time, providing a scientific basis for aquaculture decision-making. At the same time, the application of underwater robots and underwater cameras greatly improves the efficiency and accuracy of net cage inspection, reduces labor costs and safety risks. The early warning module can give early warnings of risk events such as typhoons and red tides, winning precious time for farmers to respond and effectively ensuring aquaculture safety. In addition, the remote communication and visual human-machine interface functions of the system make aquaculture management more convenient and efficient, improving the overall management efficiency. In summary, the present invention provides strong support for the sustainable development of the seawater aquaculture industry, with significant economic and social benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only the preferred embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0014] Figure 1 FIG. is a schematic diagram of the overall structure of a seawater aquaculture net cage control system proposed in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0015] The following describes the principles and features of the present invention with reference to the accompanying drawings. The listed embodiments are only used to explain the present invention and are not intended to limit the scope of the present invention.

[0016] Refer to Figure 1 , the present invention provides a seawater aquaculture net cage control system, including: A multi-parameter sensor module is deployed inside and around the cage to monitor water temperature, salinity, dissolved oxygen, pH value, ammonia nitrogen concentration, turbidity, and water flow velocity in real time, and transmit the collected data to the central control module via wireless communication; An underwater robot module is equipped with a high-definition camera, a robotic arm, and environmental sensors, and is used for autonomous inspection of the cage structure integrity, cleaning of net cage attachments, fault detection, and fish behavior observation. The movement path of the robot is dynamically optimized by the central control module based on real-time environmental data; An underwater camera module uses anti-corrosion high-definition camera equipment and is deployed in key areas of the cage to capture the activity status, feeding situation, and body surface health status of fish in real time. The image data is analyzed by AI algorithms and then fed back to the central control module; The central control module integrates an edge computing unit, which is used to receive and process multi-source monitoring data, generate optimized control strategies through built-in environmental dynamic models and fish growth models, and dynamically adjust the feeding amount, oxygenation frequency, and cage position; The data storage and analysis module constructs a distributed database to store historical environmental data, fish growth records, and equipment operation logs, and combines machine learning algorithms to mine the correlation laws between environmental parameters and aquaculture benefits to support long-term trend prediction; The early warning module, based on threshold judgment and pattern recognition technologies, conducts hierarchical early warnings for typhoon, red tide, hypoxia, and disease risk events, and pushes them to the management personnel through multiple channels including sound and light alarms, text messages, and platform pop-ups; The remote communication module uses dual links of 4G / 5G and satellite communication for remote real-time transmission of monitoring data, control instructions, and early warning information, and supports collaborative access by shore-based monitoring centers, mobile terminals, and cloud platforms; The automated execution module includes a precise feeding machine driven by a servo motor, a variable-frequency oxygenation pump, a cage mooring adjustment device, and a net cage cleaning mechanism, and performs closed-loop control operations according to the instructions of the central control module.

[0017] Exemplarily, the multi-parameter sensor module can monitor seven key indicators such as water temperature, salinity, dissolved oxygen, pH value, ammonia nitrogen concentration, turbidity, and water flow velocity in real time and transmit them to the central control module; the underwater robot module scans the cage through a high-definition camera and a side-scan sonar in the environmental sensor. When a damaged net is found, it can repair the net, clean the algae attachments on the net to avoid blockage and affect water exchange, and observe the swimming trajectory and feeding status of the fish school to assist in judging whether the aquaculture density is reasonable; the central control module can dynamically optimize the robot inspection route according to real-time sea current data to save energy consumption. The underwater camera module can deploy anti-corrosion cameras in key areas such as the feeding area and the habitat layer to capture the details of the fish body surface, and analyze whether there are signs of diseases such as white spots and rotten fins on the fish body and the intensity of competition during feeding through image recognition algorithms to provide a basis for adjusting the feeding amount. The central control module integrates multi-source data such as sensors, robots, and cameras to establish an environmental dynamic model (such as predicting water temperature changes) and a fish growth model (such as calculating the optimal feeding curve) for dynamic regulation, that is, automatically adjusting the frequency of the aeration pump according to the water temperature and fish school activity, or combining sea current data to control the mooring device to adjust the position of the cage and avoid hypoxic water masses. The data storage and analysis module can store historical environmental parameters, feeding records, disease events, etc. in a distributed database to prevent data loss, and can also analyze through machine learning: "When the salinity exceeds 32‰ continuously for 3 days, the probability of a red tide occurrence increases by 80%"; "The non-linear relationship between the feeding frequency and the growth rate of a specific fish species"; The early warning module can classify and warn typhoon, red tide, hypoxia, and disease risk events. For typhoon warnings, the force on the cage can be simulated by combining meteorological data to recommend reinforcement or transfer in advance. For disease warnings, the fish body surface images can be recognized to mark suspected infected individuals and trigger isolation instructions, and can be notified through multiple channels such as sound and light alarms, text messages, and platform pop-ups, and synchronously pushed to the sound and light alarms on the aquaculture ship, the administrator's mobile phone text messages, and the cloud platform; the remote communication module can adopt a dual-link of 4G / 5G + satellite communication to ensure real-time data transmission even 50 kilometers offshore, and can use national secret algorithms to encrypt data to prevent aquaculture parameters from being stolen. The 4G / 5G + satellite communication dual-link can also freely switch links to avoid communication interruption; the feeder in the automated execution module can deliver bait with a gram-level accuracy according to the instructions of the central control module and be calibrated in real time through a weighing sensor. The aeration pump in the automated execution module can automatically adjust the power according to the dissolved oxygen concentration. The positioning of the cage in the automated execution module can control the anchor chain through an electric winch to make the cage actively "chase light" or "avoid pollution" with the sea current.

[0018] The multi-parameter sensor module adopts a modular design. The sensor nodes can be quickly replaced through waterproof connectors and are equipped with a self-cleaning mechanism to prevent biological attachment from affecting the measurement accuracy. The spatio-temporal interpolation algorithm for sensor data fusion reconstructs the missing values through the data of surrounding nodes in case of a single-point failure, ensuring the continuity of monitoring.

[0019] Exemplarily, the multi-parameter sensor module adopts a modular design. Each sensor node is an independent unit and is connected to the main system through standardized waterproof connectors. This design makes the installation, replacement, and maintenance of the sensors extremely convenient. For example, when a sensor node needs to be replaced due to long-term immersion or accidental damage, the maintenance personnel only need to disconnect the waterproof connector and insert the new node, without the need for complex adjustments and calibrations of the entire system. To address the impact of biological attachment on the measurement accuracy of sensors in the marine environment, the multi-parameter sensor module is also equipped with a self-cleaning mechanism. This mechanism can be automatically activated according to a preset time interval or the pollution degree of the sensor surface, and removes the attachments on the sensor surface through mechanical brushing, water flow flushing, or ultrasonic vibration, etc., thus ensuring the accuracy and reliability of the measurement data. In terms of data reliability, the multi-parameter sensor module also incorporates a spatio-temporal interpolation algorithm. The marine environment is complex and changeable, and sensor nodes may experience single-point failures due to various reasons (such as the impact and corrosion of marine organisms, etc.). Once a sensor node fails, the spatio-temporal interpolation algorithm will be immediately activated, using the data of the surrounding normally operating sensor nodes, combined with the historical data in the time dimension and the data of adjacent nodes in the space dimension, to reconstruct the missing values through a mathematical model. For example, if a temperature sensor inside the cage fails, the algorithm will analyze the temperature data of the surrounding waters during the same period and the historical data of this sensor before the failure, and comprehensively calculate a reasonable temperature value, thus ensuring the continuous monitoring of the aquaculture environment parameters and the stable operation of the system.

[0020] The underwater robot module adopts a bionic propulsion system, is equipped with a multi-degree-of-freedom robotic arm and a laser calibration device, and can autonomously identify the damaged position of the net and perform repair operations. The side-scan sonar and the optical camera carried by the robot work together to construct a three-dimensional digital twin model of the cage for the decision-making reference of the central control module.

[0021] Exemplarily, in a deep - sea aquaculture area in the ocean, the robot can adopt a propulsion method by imitating the swinging of a fish tail (imitating the swimming posture of tuna). Compared with traditional propeller propulsion, its propulsion efficiency is further improved and the noise decibel is further reduced. This design enables the robot to approach the fish school near the net cage silently for behavior observation, avoiding disturbing the aquaculture organisms. The high - definition optical camera carried by the robot, in cooperation with the laser calibration device, projects a structured grating on the surface of the netting, and establishes a three - dimensional point cloud model through the triangulation algorithm. When a damaged point with a diameter > 2 cm (system - set threshold) is detected, the laser calibrator immediately locks the damaged coordinates, and takes out a pre - fabricated repair patch (made of biodegradable polymer) from the cabin through a 6 - degree - of - freedom robotic arm. The edge of the patch is aligned with the damaged area through visual servo control, with the error controlled within ±0.5 mm. Subsequently, the hot - melt device is activated to complete seamless pressing at a temperature of 120°C; The robot adopts a collaborative working mode of "side - scan sonar + multi - spectral camera": Sonar layer: The 300 kHz side - scan sonar obtains millimeter - level geometric data of the net - cage frame and constructs an underwater structure skeleton model. Optical layer: Analyzes the biomass distribution of the attachments on the netting (such as barnacles, algae) through three - band imaging of red, blue, and green. Data fusion: Overlays the sonar point cloud and the optical texture mapping to generate a three - dimensional digital twin model containing a biological attachment heat map (update frequency: once per hour). The central control module optimizes the cleaning path planning based on this.

[0022] The central control module is built - in with a multi - objective optimization algorithm, which takes into account factors such as the fish growth stage, residual bait amount, water temperature, and sea - current speed during the bait - feeding decision - making process, and dynamically generates a feeding plan; The module integrates a reinforcement learning framework and continuously optimizes the control strategy through historical operation data and aquaculture benefit feedback.

[0023] Exemplarily, the central control module dynamically generates a precise aquaculture strategy through the multi - objective optimization algorithm and the reinforcement learning framework to maximize the aquaculture benefit. For example, when a certain deep - sea aquaculture net cage enters the high - temperature period in summer, the system monitors the following parameters: Water temperature: 28°C (higher than the optimal temperature of 25°C for fish), Sea - current speed: 0.8 m / s (a strong water current may cause the bait to disperse), Residual bait amount: 15% (higher than the threshold of 10%), Fish growth stage: Adult fattening period (requiring high - protein feed). Then the algorithm decision - making process is as follows: Target weight adjustment: Prioritize reducing the residual bait amount (to avoid water quality deterioration) → Reduce the single - time bait - feeding amount by 10% and compensate for the high - temperature metabolism demand → Increase the feeding frequency by 20% to resist the influence of the water current → Select sinking - type slow - release bait. Dynamic plan generation: Original plan: Feed 3 times a day, 2 kg each time. Optimized plan: Feed 4 times a day, 1.5 kg each time, and activate the anti - water - current bait dispenser; The benefits generated by using the algorithm - decision plan: The residual bait rate drops to 8%, the feed utilization rate is increased by 12%, and the fish weight - gain speed is increased by 18%.

[0024] In the typhoon warning subsystem of the warning module, meteorological forecast data and the dynamic model of the cage structure are integrated to calculate the force distribution of the cage under different wind and wave levels, and the reinforcement or disaster avoidance plan is triggered in advance; Disease warning analyzes the characteristics of fish body surface images through a convolutional neural network, identifies common diseases such as white spot disease and fin rot disease, and marks suspected infected individuals.

[0025] For example, a typhoon warning issued by the meteorological department is received in a deep-sea aquaculture area. The central wind speed reaches level 14 (41.5m / s), and it is expected to affect the cage area in 48 hours. The typhoon warning subsystem accesses the typhoon path, wind speed, and wave height data (such as wave height>6m, period>12s) of the National Marine Forecasting Center in real time through the remote communication module, calls the cage structure dynamics model built into the central control module, inputs the cage size (length×width×depth=20m×20m×8m), material parameters (HDPE net elastic modulus 800MPa, frame yield strength 235MPa) and current anchoring status (8-point anchor chain, pre-tension 100kN), and performs simulation calculations. Based on the CFD (computational fluid dynamics) model The stress distribution of the cage frame under the coupling of typhoon waves and currents shows that the peak tension of the anchor chain in the northeast corner reaches 185kN (exceeding the safety threshold of 150kN), and the maximum deformation of the net reaches 1.2m (safety threshold <0.8m). The typhoon warning subsystem determines that the risk level is a red warning (corresponding to the combination of typhoon + extreme sea conditions) and automatically triggers the plan. The execution of the plan can be specifically as follows: the central control module sends instructions to the automation execution module: anchor adjustment device: start the electric winch to tighten the anchor chain to a pre-tension of 120kN, cage position adjustment: combined with the current forecast data, the winch differential control is used to offset the cage 50m to the southwest to the leeward waters, and multi-channel alarm: synchronously push sound and light alarms + text messages to management personnel +Platform pop-up window prompts "Red warning: It is recommended that personnel evacuate and start remote monitoring mode"; For the disease warning subsystem, for example, the underwater camera module monitors that the average weight of a fish group in a cage is 300g, clustering and out-of-group behavior occurs, and white spots appear on the body surface, then image acquisition and preprocessing can be performed. The underwater camera shoots 4K video at a frame rate of 5fps, and the AI ​​algorithm captures abnormal individual images (resolution ≥1mm / pixel). Image enhancement: The dark light enhancement algorithm (histogram equalization + Retinex filtering) is used to enhance the surface details of the fish body. Disease identification and positioning: Based on the convolutional neural network (ResNet-50 architecture, pre-trained on a database of 100,000 fish diseases) to analyze image features: white spots Disease: 0.3-0.8mm white nodules formed by parasitic ciliates on the surface of fish were detected with a confidence level of 92%. Fin rot disease: Identify the congested and ulcerated areas on the fins (area > 1cm², accounting for 30% of the fin length). Finally, target labeling was performed: the suspected infected individuals (numbered F-12, F-17, and F-23) were selected through the YOLOv5 algorithm, and the lesion coordinates and severity labels were generated for graded warnings: It was determined to be an orange warning (medium transmission risk), and the recommendation was pushed: "The sick individuals need to be isolated, the entire tank needs to be disinfected, and the salinity needs to be adjusted to 28‰ to inhibit pathogens." Automated intervention: The central control module links the feeding machine to suspend feeding and starts the aeration pump to high-frequency mode (dissolved oxygen ≥ 8mg / L) to avoid hypoxia that accelerates the disease.

[0026] The remote communication module adopts the LoRaWAN protocol to form an underwater sensor ad hoc network, and the relay buoy is equipped with a Beidou short message terminal to achieve offshore communication guarantee; The data transmission process is encrypted using the national secret SM4 algorithm, and the communication link has heartbeat detection and automatic switching functions.

[0027] For example, the underwater sensor ad hoc network consists of 10 multi-parameter sensor nodes forming a star network through the LoRaWAN protocol, and the relay buoy forwards: the surface buoy is equipped with a LoRaWAN gateway, which receives the underwater sensor data and transmits it in parallel through dual links: Main link: 4G LTE-Advanced Pro, backup link: BeiDou-3 RDSS short message; national secret SM4 encryption: original data (such as water temperature 23.5℃, dissolved oxygen 6.8mg / L) is encrypted by SM4 algorithm (key length 128bit, CBC mode), and ciphertext is generated with MAC check code (32bit). For link priority, full data is transmitted through 4G link by default (100 groups per second), and BeiDou link only transmits key parameters (1 heartbeat packet + abnormal data every 5 minutes). When the typhoon causes the 4G base station to lose power, the communication module detects that the 4G link packet loss rate is greater than 95% (for 3 minutes), and automatically triggers link switching: BeiDou short message is upgraded to the main link, and sensitive parameters such as dissolved oxygen and pH value are transmitted first. The message content is compressed to 80 bytes (using LZ4 algorithm), and the shore-based center receives it synchronously: key data monitoring is restored through the BeiDou command user terminal to ensure that the control instructions of the oxygen pump are not interrupted.

[0028] The bait throwing machine of the automation execution module is equipped with a weighing sensor and an image recognition unit to calibrate the amount of bait thrown and count the feeding activity in real time; The cage position adjustment device controls the retraction and release of the anchor chain through an electric winch, and actively avoids severe water masses based on ocean current forecast data.

[0029] For example, for the changes in fish activity and feeding under high temperature conditions, when the underwater camera analyzes the fish distribution density (0.3 fish / m², a decrease of 40%) and feeding frequency (1.2 pecking times per second, a decrease of 35%) through the AI ​​algorithm, the central control module calls the fish growth model to calculate the basal metabolic rate at the current water temperature (an increase of 15% compared to normal temperature), and generates a feeding plan based on the amount of residual bait (historical release amount × residual coefficient 0.1): from the original plan of 8kg / time → adjusted to 6.5kg / time, the image recognition unit, by photographing the feeding area, counts the feeding activity (by tracking the movement trajectory of the fish body), and feeds back the actual feeding rate of 78%. Then, the central control module adjusts the next feeding amount to 6.8kg based on the feeding rate data, and links the aeration pump to increase the frequency (original 20min / time → 15min / time) to alleviate high temperature stress.

[0030] When the early warning module triggers a yellow warning through threshold judgment and pushes disaster avoidance suggestions, the central control module calls the ocean current numerical model to predict the movement path of the water mass (speed 0.6 m / s, direction northeast). Then, combined with the mooring state of the cage (current position at point A, water depth 25 m), it plans the migration path to point B (water depth 20 m, 300 m away from point A). Then the electric winch retracts and releases the anchor chain step by step (the northwest anchor chain is shortened by 10 m, and the southeast anchor chain is extended by 15 m), and controls the cage to move southeast at a speed of 0.3 m / min to avoid the adverse water mass.

[0031] This embodiment includes a visual human-machine interface, which displays the distribution of cages, the heat map of environmental parameters, and the operating status of equipment on a 3D GIS map; The visual human-machine interface integrates a virtual reality (VR) operation mode, supporting remote first-person perspective inspection and manual intervention of key equipment. Exemplarily, when managers need to monitor the environment and operating status of multiple cages, they can view the cage distribution, the heat map of environmental parameters, and the operating status of equipment through the 3D GIS map. The visual human-machine interface uses a satellite map as the base map, superimposing the 3D model of the cage (distributed according to actual coordinates), supporting zooming / rotation / section view switching; when clicking on a single cage, a floating window pops up to display real-time parameters: Environmental heat map: water temperature (gradual change from red to blue, current 26.8 °C), dissolved oxygen (gradual change from green to yellow, 7.2 mg / L), and the equipment status can display the aerator (operating / green), the feeder (standby / grey), and the mooring system (normal / blue). When the turbidity in a certain cage area suddenly increases to 50 NTU (threshold 30 NTU), the map automatically highlights this area and flashes a red border, and simultaneously pushes a pop-up window: "Abnormal turbidity around the cage, it is recommended to start an underwater robot to investigate the pollution source"; when managers need to remotely inspect the cage structure and intervene in equipment failures, they can use the VR inspection mode: After wearing the VR helmet, the interface switches to the first-person perspective, "enters" the inside of the cage through the real-time images of the underwater robot / camera, and intervenes through the following operations: Gesture interaction: Wave the hand to switch the perspective (top view / side view / fish school tracking mode), Pinch gesture to zoom in and observe the details of the fish body (such as body surface damage), Voice command: "Detect net damage", and the AI automatically marks the suspected leak location.

[0032] Remote equipment intervention: Manipulator control: After discovering a torn net, the manager "grabs" the virtual joystick in the VR interface and remotely controls the manipulator of the underwater robot for repair (laser positioning error < 5 mm).

[0033] Parameter adjustment: Through the virtual control panel, manually adjust the feeding amount (original 200 g / time → 150 g / time), and the interface displays the remaining bait amount and the feedback of the feeding activity of the fish school in real time.

[0034] This system realizes the intelligentization and remote control of aquaculture management through 3D visualization, VR interaction and multi-terminal collaboration, which can significantly improve the monitoring efficiency and emergency response capabilities.

[0035] The above are only the preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A marine aquaculture cage management and control system, characterized in that: include: Multi-parameter sensor modules are deployed inside the cage and in the surrounding waters to monitor water temperature, salinity, dissolved oxygen, pH value, ammonia nitrogen concentration, turbidity and water flow rate in real time, and transmit the collected data to the central control module via wireless communication; The underwater robot module is equipped with a high-definition camera, a robotic arm and environmental sensors, which are used for autonomous inspection of the structural integrity of the cages, cleaning of net attachments, troubleshooting and fish behavior observation. The robot's motion path is dynamically optimized by the central control module based on real-time environmental data; The underwater camera module uses anti-corrosion high-definition camera equipment and is deployed in key areas of the cage to capture fish activity, feeding conditions, and body health in real time. The image data is analyzed by the AI ​​algorithm and fed back to the central control module. The central control module integrates an edge computing unit to receive and process multi-source monitoring data, generate an optimization control strategy through the built-in environmental dynamic model and fish growth model, and dynamically adjust the feeding amount, oxygenation frequency and cage position; The data storage and analysis module builds a distributed database to store historical environmental data, fish growth records, and equipment operation logs. It uses machine learning algorithms to mine the correlation between environmental parameters and aquaculture benefits, and supports long-term trend prediction. The early warning module, based on threshold determination and pattern recognition technology, provides graded early warnings for typhoon, red tide, hypoxia, and disease risk events, and pushes them to management personnel through multiple channels such as sound and light alarms, text messages, and platform pop-up windows; Remote communication module, which uses 4G / 5G and satellite communication dual-link to monitor data, control instructions and early warning information for remote real-time transmission, and supports collaborative access by shore-based monitoring centers, mobile terminals and cloud platforms; The automated execution module, including a servo motor-driven precision bait-casting machine, a variable frequency oxygen pump, a cage anchoring adjustment device, and a net cleaning mechanism, performs closed-loop control operations according to instructions from the central control module.

2. The marine aquaculture cage management and control system according to claim 1, characterized in that: The multi-parameter sensor module adopts a modular design, the sensor nodes are quickly replaced through waterproof connectors, and is equipped with a self-cleaning mechanism to prevent biological attachment from affecting measurement accuracy; The sensor data is fused with a spatiotemporal interpolation algorithm to reconstruct missing values ​​through surrounding node data in the event of a single point failure, ensuring monitoring continuity.

3. The marine aquaculture cage management and control system according to claim 1, characterized in that: The underwater robot module adopts a bionic propulsion system, is equipped with a multi-degree-of-freedom robotic arm and a laser calibration device, and can autonomously identify the damaged position of the net and perform repair operations; The side-scan sonar and optical camera carried by the robot work together to build a three-dimensional digital twin model of the cage for decision-making reference by the central control module.

4. The marine aquaculture cage management and control system according to claim 1, characterized in that: The central control module has a built-in multi-objective optimization algorithm, which takes into account the fish growth stage, residual bait amount, water temperature and current speed in the decision-making of bait delivery, and dynamically generates a feeding plan; The module integrates a reinforcement learning framework to continuously optimize the control strategy through historical operation data and breeding benefit feedback.

5. The marine aquaculture cage management and control system according to claim 1, characterized in that: In the typhoon warning subsystem of the warning module, the weather forecast data and the cage structure dynamics model are integrated to calculate the force distribution of the cage under different wind and wave levels, and trigger the reinforcement or disaster avoidance plan in advance; Disease warning uses convolutional neural networks to analyze fish surface image features, identify common symptoms of white spot disease and fin rot, and mark suspected infected individuals.

6. The marine aquaculture cage management and control system according to claim 1, characterized in that: The remote communication module adopts the LoRaWAN protocol to form an underwater sensor ad hoc network, and the relay buoy is equipped with a Beidou short message terminal to achieve offshore communication guarantee; The data transmission process is encrypted using the national secret SM4 algorithm, and the communication link has heartbeat detection and automatic switching functions.

7. The marine aquaculture cage management and control system according to claim 1, characterized in that: The bait throwing machine of the automation execution module is equipped with a weighing sensor and an image recognition unit to calibrate the amount of bait thrown and count the feeding activity in real time; The cage position adjustment device controls the retraction and release of the anchor chain through an electric winch, and actively avoids severe water masses based on ocean current forecast data.

8. The marine aquaculture cage management and control system according to claim 1, characterized in that: Includes a visual human-machine interface that uses a three-dimensional GIS map to display cage distribution, environmental parameter thermal maps, and equipment operating status; The visual human-machine interface integrates virtual reality (VR) operation mode, supporting remote first-person inspection and manual intervention of key equipment.

Images