Intelligent fishery breeding device and system

Abstract

The invention relates to the technical field of fishery breeding, in particular to an intelligent fishery breeding device and system, which adopts a vertical integrated framework of a breeding layer, a fish collecting layer and a water storage layer. The two groups of fish tanks on the culture layer are in back symmetry to form a 50 mm biological behavior observation channel, and the water quality and fish school behaviors are monitored in real time in combination with a sensor group and a camera; the fish collecting layer realizes fish school gravity sorting through an electric ball valve and an angle-adjustable PVC (Polyvinyl Chloride) pipeline; waste water purified by the water storage layer three-stage filtering device is recycled in a closed loop through a water suction pump. The system is based on linkage regulation and control of a central control module, wherein when ammonia nitrogen is larger than 1.5 mg / L, water body updating is accelerated, automatic sorting is triggered by abnormal behaviors of fish schools, and the feeding amount is dynamically optimized through residual feed weighing. And the maintenance efficiency is remarkably improved by the modularized quick-release filter unit and the gravity fish unloading mechanism. The problems of low space utilization rate, water quality deterioration, disease response lag and resource waste of traditional breeding are solved.

Description

Technical Field

[0001] This invention relates to the field of aquaculture technology, specifically to an intelligent aquaculture device and system. Background Technology

[0002] Traditional aquaculture has long faced the problem of low space utilization. Common planar aquaculture systems require large areas of water or land, which not only leads to land resource scarcity but also limits the stocking density per unit area. This extensive layout makes water quality management difficult, and local water bodies are prone to deterioration due to the accumulation of excrement, which in turn induces frequent outbreaks of fish diseases. Aquaculture farmers often rely on manual inspections to observe the condition of fish schools, making it difficult to achieve early detection and precise intervention of diseases.

[0003] In terms of aquaculture process control, existing facilities generally suffer from insufficient automation. Key aspects such as water temperature control, feeding operations, and water quality monitoring still heavily rely on manual experience, which is not only labor-intensive but also results in poor control effects due to lag in response. Especially in handling abnormal situations, such as fish disease outbreaks or sudden changes in water quality, traditional methods lack rapid and effective response mechanisms, often leading to economic losses due to untimely handling.

[0004] Water waste also hinders the industry's sustainable development. Open-type aquaculture systems require a continuous influx of fresh water to maintain water quality, and the direct discharge of large amounts of nutrient-rich aquaculture wastewater exacerbates water consumption and causes ecological pollution. Although some facilities have attempted to incorporate water recycling designs, the actual reuse rate is insufficient to meet the needs of intensive aquaculture due to low filtration efficiency and inadequate system integration.

[0005] Furthermore, existing fishery equipment generally lacks intelligent decision support capabilities. A closed-loop link has not been formed between environmental data collection and the implementation of control measures by agencies, making it difficult for fish farmers to obtain real-time early warnings and scientific guidance. This disconnect between perception and decision-making makes it difficult to implement refined aquaculture management and hinders the industry's upgrade towards higher efficiency and intelligence. Summary of the Invention

[0006] The purpose of this invention is to provide an intelligent aquaculture device and system that achieves high-density aquaculture, dynamic water quality control, precise disease sorting, and efficient resource utilization through a vertical three-dimensional architecture, closed-loop water circulation, machine vision monitoring, and modular quick-disassembly design.

[0007] To achieve the above-mentioned technical objectives and effects, the present invention is implemented through the following technical solution: A smart aquaculture device includes a supporting frame, which has an aquaculture layer, a fish collection layer, and a water storage layer arranged from top to bottom. In the aquaculture layer, two polycarbonate fish tanks are arranged symmetrically back to back, with a 50mm gap to form a biological behavior observation channel to reduce camera blind spots; each fish tank has a sensor group on its side wall, and a water replenishment pool and its water replenishment pipeline are installed on the top. The side wall of the fish tank is connected to the feeding mechanism to achieve precise feeding. The electric ball valve input end of the fish collection layer is connected to the drain outlet at the bottom of the fish tank through a sealing flange. The electric ball valve output end is equipped with a clamp to fix an adjustable angle PVC pipe. The lower end of the PVC pipe hangs down towards the fish truck with a net. The trapdoor at the rear of the fish truck is connected to the fish truck body through a spring hinge. The trapdoor is locked on both sides with a pin-type snap fastener, forming a gravity unloading fish mechanism. The water storage layer receives the seepage water from the fish collection layer. The three-stage filtration device is fixed by a PP cotton layer, an activated carbon layer, and an ultrafiltration membrane layer from top to bottom through a chute. This is used to remove impurities, adsorb organic matter, and purify microorganisms. After filtration, the water flows into the water storage tank through a guide channel and is then pumped into the aquaculture layer replenishment tank by a water pump through a UPVC pipe, forming a closed water cycle. The supporting frame is assembled into a three-dimensional truss from 20-series aluminum profiles using T-slot bolts. The aquaculture layer, fish collection layer, and water storage layer are positioned at a preset elevation using angle brackets. The frame is machined by five-axis CNC to ensure the straightness tolerance of the sensor mounting surface is ±0.1mm / m, and the surface is anodized to resist corrosion in high-humidity environments.

[0008] On the other hand, the present invention proposes a system based on the above-mentioned device, including: an environmental control module, a feeding control module, a health monitoring module, a water circulation control module, and a central control module; the central control module is connected via electrical lines to a sensor group bolted to the side wall of the aquarium in the breeding layer, a camera aimed at the biological behavior observation channel, a servo motor of the top feeding device, an electric ball valve at the bottom drain of the aquarium, and a water pump in the water storage layer; the central control module generates control commands to drive the start and stop actions of the servo motor, the electric ball valve, and the water pump based on the water quality data collected by the sensor group and the fish behavior data captured by the camera.

[0009] Furthermore, the environmental control module reads data from the sensor group installed on the sidewalls of the two polycarbonate aquariums in the aquaculture layer in real time, including water temperature, pH value, dissolved oxygen, and ammonia nitrogen concentration; and adjusts the water inflow of the two aquariums respectively through independent electrically controlled valves connected to the top flange of the compartmentalized water replenishment pool; when the ammonia nitrogen concentration of a certain aquarium is detected to exceed the set threshold of 1.5mg, the opening of the corresponding electrically controlled valve of that aquarium is increased to accelerate water renewal.

[0010] Furthermore, the feeding control module controls the rotation angle of the MG996R servo motor in the feeding device snapped onto the top of the aquaculture layer; based on the fish distribution density captured by the camera in the biological behavior observation channel, the feeding frequency is calculated. When the percentage of pixels in the observation channel area where fish gather exceeds 70%, the servo motor is triggered to open the gate at 10-second intervals. When the percentage of pixels when the fish are dispersed is less than 30%, stop feeding.

[0011] Furthermore, the health monitoring module captures video streams of fish swimming at 30fps using a camera directly facing the fish through a 50mm wide observation channel; it analyzes the frequency of fish rubbing against the tank wall and the duration of surfacing: if the rubbing frequency is >3 times / minute within 5 consecutive minutes, it is marked as abnormal behavior; if the surfacing duration is >2 minutes, a first-level warning signal is sent to the central control module; when the abnormality lasts for 5 minutes, the corresponding bottom ball valve of the fish tank is opened, allowing the fish to fall into the stainless steel woven mesh of the mesh-type fish cart through an adjustable-angle PVC pipe.

[0012] Furthermore, the water circulation control module monitors the operating status of the three-stage filtration device in the water storage layer, including: detecting turbidity changes through a water quality sensor at the outlet of the guide channel; determining that the ultrafiltration membrane layer is clogged when the turbidity value increases by 40% compared to the initial value; and controlling the water pump to pump the filtered water from the water storage tank into the aquaculture layer replenishment tank through the UPVC pipeline to maintain a closed water circulation.

[0013] Furthermore, it also includes a fish disease detection module, including: MaixCam cameras, deployed in the biological behavior observation channel of the aquaculture layer, are used to capture real-time video streams of fish schools; The image processing unit, which communicates with the camera, has a built-in visual model trained based on the YOLOv5 algorithm, used to identify fish disease characteristics in the video stream. The server receives disease detection result data output by the image processing unit; The client-side interface displays real-time camera footage and labels the identified disease types and confidence levels, while generating historical records including timestamps, disease types, and severity. The visual model performs the following: real-time analysis of video frames captured by the camera; identification of diseased fish and output of disease type and confidence level; the client interface triggers a system warning when the confidence level exceeds a threshold, and supports querying historical records to generate disease statistical reports.

[0014] Furthermore, the feeding control module dynamically adjusts the feeding amount based on the weight of the uneaten feed collected at the bottom of the fish collection layer net-type fish truck: after the daily feeding is completed, the weight of the uneaten feed recorded by the weighing sensor is read; if the weight of the uneaten feed accounts for more than 15% of the total feeding amount for the day, the basic feeding amount for the next day is reduced proportionally; if the weight of the uneaten feed is less than 5%, the feeding amount for the next day is increased by 10%.

[0015] Furthermore, the central control module processes the dissolved oxygen data from the sensor group: when the dissolved oxygen concentration is <4mg / L, it immediately starts the oxygenation equipment and opens the water supply valve; when the ammonia nitrogen concentration is >1.5mg / L, it accelerates the operation frequency of the water pump to 120% of its rated power to enhance the water circulation filtration efficiency.

[0016] Furthermore, water level sensors are embedded in the bottom of the two independent chambers of the compartmentalized water replenishment tank; the water level data of each chamber independently controls the corresponding fish tank's water replenishment electronic control valve. When the water level is 20mm below the calibrated height, the electrically controlled valve will be opened to replenish water. When the water level reaches the set height, the electric control valve is closed and a water replenishment completion signal is sent.

[0017] The beneficial effects of this invention are: This invention utilizes a vertically stacked design of aquaculture, fish collection, and water storage layers. A 20-series aluminum profile frame is assembled with T-slot bolts and precision-machined using five-axis CNC machining, achieving a synergistic optimization of structural rigidity and space compression. The back-to-back symmetrical arrangement of the polycarbonate aquariums forms a 50mm biological behavior observation channel, eliminating blind spots for cameras while reducing lateral footprint. The fish collection layer features an electric ball valve and adjustable-angle PVC pipe directly connected to the bottom drain of the aquarium, using gravitational potential energy to precisely guide the fish into the mesh-type fish cart. The water storage layer features a tiered, three-stage filtration system beneath the fish cart, achieving efficient collection and purification of leaked water. Through space reuse and a gravity-driven mechanism, this invention completely solves the problem of low land utilization in traditional planar aquaculture.

[0018] This invention constructs a complete water circulation chain based on the graded purification mechanism of a three-stage filtration device and the dynamic regulation of a central control module: the synergistic effect of the PP cotton layer intercepting solid impurities, the activated carbon layer adsorbing dissolved organic matter, and the ultrafiltration membrane layer retaining microorganisms allows the filtered water to flow into the storage tank via a guide channel; the water pump pumps the purified water through UPVC pipes into the aquaculture replenishment tank to form a closed loop. Water level sensors embedded at the bottom of the compartmentalized replenishment tank independently control the water replenishment valves of each aquarium, automatically opening and closing when the water level deviates from the calibrated value by 20mm; the central control module processes dissolved oxygen <4mg / L or ammonia nitrogen >1.5mg / L in real time, simultaneously triggering the oxygenation equipment, water replenishment valves, and water pumps to overclock to 120% of their rated power. By enhancing water renewal and filtration efficiency, this invention breaks the water quality deterioration chain, fundamentally solving the problems of water resource waste and pollution discharge in traditional aquaculture.

[0019] This invention utilizes a Maix-CAM camera and a group of sensors on the sidewalls of the aquarium within a biological behavior observation channel to construct a data-driven decision-making system: when the camera captures fish rubbing against the tank wall more than 3 times per minute or surfacing for more than 2 minutes, it automatically opens the corresponding electric ball valve at the bottom of the aquarium, allowing the fish to fall into the fish cart through an adjustable-angle PVC pipe for rapid sorting; based on the pixel ratio of fish distribution in the observation channel, it controls the rotation angle of the MG996R servo motor to achieve precise feeding; and it dynamically adjusts the feeding amount for the next day by combining the weight of uneaten feed recorded by the weighing sensor at the bottom of the fish cart. The STM32F407VE microcontroller integrates and processes data from multiple sensors, linking the actuators to form a millisecond-level response closed loop, overcoming the pain points of delayed disease response and blind feeding in traditional aquaculture.

[0020] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a schematic diagram of the overall structure of the present invention; Figure 2 This is a schematic diagram of the aquaculture layer structure; Figure 3 This is a schematic diagram of the fish-collecting layer structure; Figure 4 This is a schematic diagram of the fish truck structure; Figure 5 This is a schematic diagram of the aquifer structure; Figure 6 This is a schematic diagram of the sensor structure of the present invention; Figure 7 This is a partial structural diagram of the present invention; Figure 8 This is a schematic diagram of a fish tank structure; Figure 9 This is a sample image of the app page. Figure 10 This is a diagram illustrating the testing of normal fish. The attached diagram lists the components represented by each number as follows: 1. Aquaculture layer, 2. Fish collection layer, 3. Water storage layer, 4. Feeding mechanism, 5. Camera, 6. Support frame, 101. Fish tank, 102. Water supply pipe, 103. Temperature sensor, 104. pH sensor, 105. Turbidity sensor, 201. Electric ball valve, 202. PVC pipe, 203. Fish cart, 2031. Trap door, 2032. Pin-type snap fastener, 301. Filter device, 302. Water storage tank, 303. Water supply tank, 401. Feed hopper, 402. Feed outlet, 403. Steering motor. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] Example 1 The intelligent aquaculture device described in this embodiment includes a vertically integrated three-layer structure: the aquaculture layer is provided with two polycarbonate fish tanks arranged symmetrically back to back, the two sets of fish tanks are 50mm apart to form an observation channel, each fish tank is provided with a water replenishment pool and its water replenishment pipeline on the top, and sensor brackets are installed on the side walls to fix including but not limited to temperature sensors, pH sensors and turbidity sensors. The fish collection layer is located directly below the aquaculture layer. It includes an electric ball valve with an inlet flange connected to the bottom outlet of the aquarium, an adjustable-angle PVC pipe with an inlet flange connected to the outlet of the electric ball valve, and a mesh-type fish cart that receives the outlet of the pipe. The bottom of the fish cart is a stainless steel woven mesh structure, and the rear is equipped with a hinged trapdoor and double-sided latches. Four universal casters are installed at the bottom of the cart. When the electric ball valve is energized, the fish fall vertically through the PVC pipe with the water flow. After hitting the fish cart filter, they are trapped. The latches on both sides are opened manually by rotating them. The trapdoor is opened by gravity and rotates downward around the hinge axis. The fish slide out along the inclined surface of the trapdoor. The water storage layer includes a three-layer filtration device with its inlet facing the bottom of the fish truck's mesh area, a water storage tank receiving the filter's outlet, and a centrifugal water pump connected to the water storage tank via an inlet pipe. The pump's outlet connects to a compartmentalized replenishment tank at the top of the aquaculture layer. This replenishment tank is divided into two independent chambers, each connected to the corresponding aquarium's top inlet via an independent ball valve, and a water level sensor is embedded at the bottom of each chamber. The entire structure is supported by a 20-series aluminum profile frame assembled with T-slot bolts, and key connection nodes are machined using five-axis CNC machining. The water circulation path is as follows: water leaking from the fish truck → water receiving trough at the top of the filtration device → vertically penetrating the PP cotton filter plate → flowing through the gaps in the granular activated carbon filter plate → horizontally passing through the hollow fiber ultrafiltration membrane plate → flowing into the water storage tank via a guide channel → the water pump operates to pump the water into the aquaculture layer's replenishment tank. In this embodiment, the filtration device includes three vertically stacked detachable units: the upper layer is a PP cotton filter box, the middle layer is an activated carbon filter box, and the lower layer is an ultrafiltration membrane filter box. Each filter box is fixed to the galvanized steel plate shell by quick-release clamps. The trapdoor of the net-type fish cart is kept horizontally closed by gravity when the latch is locked. After unlocking, it rotates downward around the hinge axis to open and unload the cargo.

[0025] In this embodiment, a feeding mechanism is also included: a feeding port is provided at the bottom of the storage hopper, and the MG996R servo motor drives the gate through the swing arm to control the opening and closing of the feeding port. This mechanism is fixed to the side wall of the fish tank by an L-shaped bracket; a rotating joint is provided in the middle of the adjustable angle PVC pipe to adjust the outlet tilt angle so that the fish can fall accurately into the center area of ​​the fish truck.

[0026] In this embodiment, the observation channel faces the acquisition window of the Maix-CAM camera, and the transparent sidewalls of the double fish tanks form a 300mm×200mm optical monitoring area; the aluminum profile frame nodes are machined with a straightness tolerance of ±0.1mm / m and the surface is anodized.

[0027] Example 2 Based on structural stability requirements, the main frame of this device is made of 20-series aluminum profiles (such as...). Figure 1 As shown in the figure, its topology optimization design achieves a tensile strength of ≥240MPa. T-slot bolt connections improve assembly efficiency by 300%-500% compared to traditional arc welding, avoiding dimensional deviations caused by thermal deformation. Five-axis CNC precision machining ensures a straightness tolerance of ±0.1mm / m for key nodes, meeting the requirements for precision sensor installation. Surface anodizing further enhances the structure's resistance to stress corrosion, making it more suitable for high-humidity aquaculture environments.

[0028] Aquaculture layer To monitor the fish population and ecological environment in real time, this device employs a dual-water-body isolated aquaculture unit design, such as... Figure 2 As shown, two polycarbonate (PC) aquariums are arranged symmetrically back-to-back, with a 50mm gap between them to form a biological behavior observation channel, reducing the blind spot for video capture of fish behavior by >60%. A sensor system (temperature sensor, pH sensor, turbidity sensor) is placed on the sides of each aquarium. Figure 8 As shown, this system monitors the water temperature, pH level, and turbidity, providing real-time feedback to the backend. Above the aquaculture layer, a pool is placed to store water drawn from the reservoir at the bottom of the device. Two electric ball valves are installed above the two tanks to replenish the water. Additionally, feeding devices are added to the sides of both tanks, allowing for precise feeding via settings or user control.

[0029] Fish gathering layer To achieve rapid sorting, the device is equipped with a fish collection layer, such as... Figure 3 As shown, the fish collection layer of the device consists of two fish tanks at the top, two electric ball valves, two adjustable-angle PVC pipes, and two fish carts. The electric ball valves are connected to the fish tanks at the top and to an adjustable-angle PVC pipe at the bottom. By controlling the electric ball valves, the fish are collected; when the valves open, the fish fall through the pipes into the fish carts. The fish carts use a mesh-like structure, consisting of a cart body, handlebars, trapdoors, pulleys, and two latches. To facilitate unloading the fish, a trapdoor is installed at the rear of the cart. During unloading, the latches on both sides are opened, and the fish fall due to gravity. This reduces the time for a single fish collection operation to two-thirds of traditional methods, making it suitable for high-density recirculating aquaculture systems.

[0030] aquifer Adopting a modular water circulation architecture, such as Figure 4 As shown, the water collected from the fish tank above is collected by gravity and flows through the fish cart to the filtration system. The water then passes through a three-layer filtration system (PP cotton, activated carbon, and ultrafiltration membrane) and is collected in a storage tank. A water pump then pumps the water from the storage tank to a pool above the fish tank, achieving water recycling.

[0031] Hardware system implementation This system consists of six modules: a one-click harvesting module, a control module, a water resource recycling module, a data acquisition module, a communication module, and a platform and terminal interaction module. Each module is designed according to the actual needs of the system, and the specific module design ideas are described below.

[0032] One-click fishing module Traditional seine net fishing is energy-intensive and highly dependent on manual labor. The emergence of gravity-fed fish collection systems has changed the game—by pressing a button on the control panel, a bottom solenoid valve automatically opens, creating a water flow funnel, allowing fish to naturally flow into the fish truck. This effectively reduces energy waste and labor costs, making it suitable for high-density aquaculture ponds with rapid turnover. It improves the utilization rate of fishery resources and the level of automation, providing strong technical support for the sustainable development of fisheries.

[0033] Control module In the design of the intelligent fish farming control system, the STM32F407VE microcontroller is used as the main control core. It features low power consumption, high speed, and abundant resources, enabling precise acquisition of environmental parameters such as water temperature, pH value, and turbidity. It also supports simultaneous connection of multiple devices, achieving efficient data communication. Its ample GPIO ports can simultaneously drive multiple actuators such as feeding equipment, aerators, and water circulation systems, ensuring the stability of the farming environment. The internal timer can generate PWM waveforms of different frequencies to drive and control the relay module, improving the system's stability and accuracy, and meeting the needs of fish growth and reproduction.

[0034] Data acquisition module The data acquisition module includes image data acquisition and various sensor data acquisition. In this study, a Maix-CAM high-definition camera module (such as...) was used. Figure 7 The system collects image data and uses the YOLOv5 algorithm to identify the health status of fish, thereby achieving informatized, automated, and intelligent management of the fish farming process. The image acquisition module is responsible for acquiring, analyzing, processing, integrating, and identifying underwater behavioral characteristics of the fish, providing support for subsequent image fusion analysis, judgment, evaluation, and prediction by computer software. In addition, the system is equipped with DS18B20 temperature sensors, pH sensors, and turbidity sensors to ensure the completion of data acquisition and transmission.

[0035] In this embodiment, the DS18B20 temperature sensor was used for monitoring the water temperature in the fishpond. It is easy to wire, has low latency, strong anti-interference capabilities, and transmits data via digital signals. After acquiring water temperature data, accurate water temperature information is obtained by reading and writing to the sensor's internal registers, offering advantages such as low cost and simple programming.

[0036] For pH monitoring, this embodiment uses a pH electrode-based sensor. Its sensitive glass membrane can react with hydrogen ions to form a potential difference between positive and negative ions, generating a weak electrical signal. When the potential difference between the reference electrode and the pH electrode remains constant, this electrical signal can be accurately converted into a specific pH value, thus achieving accurate water quality detection.

[0037] In this embodiment, a turbidity sensor is used to detect water turbidity. The core of this sensor is a pair of infrared emitting and receiving tubes, and its working principle is based on the influence of water turbidity on light transmittance. Specifically, the higher the water turbidity, the less light transmits. The light receiver converts the received light intensity into a corresponding current signal; that is, the more light transmits, the stronger the current signal; conversely, the less light transmits, the weaker the current signal. This allows for the determination of whether the fishpond needs a water change to maintain a suitable aquaculture environment.

[0038] Communication module In the field of contemporary wireless communication technology research, the ESP32 WiFi module is widely adopted due to its superior functionality and high performance. The ESP32 module not only supports WiFi connectivity but also integrates Bluetooth technology, making it a preferred core component for IoT applications. In this research project, the ESP32 WiFi module was selected as the core component of the wireless communication system.

[0039] The STM32 microcontroller, with its superior performance and low power consumption, holds a pivotal position in the field of embedded systems. In this research design, the STM32 (REF _Ref23742 \w \h) establishes a connection with the ESP32 module through its built-in serial communication interface. The STM32 is responsible for acquiring various sensor data and transmitting these data to the ESP32 module via the serial interface.

[0040] After receiving data from the STM32 microcontroller, the ESP32 module utilizes its built-in network capabilities to transmit the data to a remote cloud server in real time via the WebSocket network communication protocol. Based on this, an efficient and stable wireless data transmission system is built, ensuring that data collected by the STM32 microcontroller can be transmitted to the cloud server quickly and accurately, providing a solid foundation for subsequent data analysis and processing.

[0041] Platform and terminal interaction module This study utilizes UniAPP technology to develop a mobile terminal, achieving efficient data interaction with the server through the TCP network communication protocol. This module is responsible for recording and displaying key information about the fishponds, enabling real-time monitoring and intelligent control of the fishpond environment. After entering the control interface for each fishpond, users can view real-time key data such as the current pH value, temperature, and oxygen content, and adjust the environmental parameters in the fishpond by setting target values. Furthermore, when abnormal conditions occur in the fishpond, the system will immediately activate an early warning mechanism and display a prompt message in the application to remind users to take timely measures to ensure the health of the fish and the safety of the aquaculture environment.

[0042] (1) APP homepage: The top pie chart displays the distribution of fishpond status. The pie chart is divided into three parts: Normal (green): Number of fishponds currently in a normal state; Abnormal (red): Number of fishponds exhibiting abnormal conditions; Offline (gray): Number of fishponds that are offline. Numerical statistics: On the right side of the pie chart, the number of fishponds in each status category is listed in detail numerically for quick understanding of the overall situation. Below the chart is a task management area, displaying the number of abnormal tasks and scheduled tasks. Clicking on these tasks will take you to the task management page to view and process specific tasks. The middle section is the fishpond management area, listing basic information for all fishponds, including key parameters such as fishpond name, pH value, and oxygen content, along with thumbnails of the fishponds for easy identification.

[0043] (2) Category Page: A search box is located at the top of the page, allowing users to quickly search for the fish or equipment name and find the information they need. The middle section displays brief introductions to various fish and equipment, each accompanied by a picture and a short description to help users quickly understand their characteristics and uses. Clicking on these entries will take you to a detailed page for more in-depth information.

[0044] (3) AI Page: The AI-powered page is a smart diagnostic Q&A platform specifically designed for fish disease detection. Users can ask the AI ​​questions or upload images, which the AI ​​will analyze and provide the causes and corresponding solutions. The page also includes quick access to frequently asked questions, such as "What should I do if my fish gets white spot disease?", allowing users to quickly obtain information. Through intelligent diagnosis, the AI-powered page provides users with scientific and efficient disease prevention and control guidance, helping to ensure smooth aquaculture operations.

[0045] Water resource recycling module Addressing the water waste and pollution issues inherent in traditional aquaculture methods, this system aims to create a "closed-loop, zero-emission" aquaculture environment. It employs a four-stage process chain: physical filtration removes solid impurities from the water, biological purification treats organic matter, disinfection ensures water quality safety, and finally, a pump returns the treated water to the fish tank, achieving water recycling. This design not only embodies technological innovation but also integrates intelligent control technology, providing a sustainable water resource management solution. This effectively alleviates the conflict between industrial development and environmental protection and provides technical support for the green transformation and upgrading of the aquaculture industry. The attached diagram shows the actual water pump, high-definition camera, and filtration device structure.

[0046] Fish disease detection module This includes using a MaixCam camera combined with YOLOv5 technology to train a highly efficient visual model, enabling intelligent identification and real-time monitoring of fish diseases. This module uses the efficient visual model to perform accurate and real-time detection and analysis of fish diseases. For example... Figure 10 As shown, users can view real-time camera feeds and fish health status monitoring results on the client side. MaixCam continuously monitors the fishpond, using deployed models for real-time analysis to quickly identify disease characteristics. Detection results are transmitted to the server via the network and pushed to the user's client interface. The system labels the disease type and confidence level on the screen. Furthermore, the system has a historical record function, including timestamps, disease types, and severity information, facilitating user queries, statistical report generation, and disease trend analysis.

[0047] In summary, this invention proposes an intelligent aquaculture device and system, employing a vertically integrated architecture comprising an aquaculture layer, a fish collection layer, and a water storage layer. The aquaculture layer features two symmetrically arranged aquariums facing away from each other, forming a 50mm biological behavior observation channel. Combined with a sensor array and camera, it monitors water quality and fish behavior in real time. The fish collection layer uses electric ball valves and adjustable-angle PVC pipes to achieve gravity-based sorting of fish. The water storage layer employs a three-stage filtration system to purify wastewater, which is then recycled in a closed-loop system via a pump. The system is based on a central control module for coordinated regulation: water renewal is accelerated when ammonia nitrogen > 1.5 mg / L; abnormal fish behavior triggers automatic sorting; and uneaten feed is weighed to dynamically optimize feeding amounts. Modular, quick-release filter units and a gravity-operated fish unloading mechanism significantly improve maintenance efficiency. This invention solves the problems of low space utilization, water quality deterioration, delayed disease response, and resource waste inherent in traditional aquaculture.

[0048] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. An intelligent aquaculture device, characterized in that, It includes a supporting frame (6), and from top to bottom, it is provided with an aquaculture layer (1), a fish collection layer (2), and a water storage layer (3). In the aquaculture layer (1), two polycarbonate fish tanks (101) are arranged symmetrically back to back, with a spacing of 50mm to form a biological behavior observation channel to reduce the blind spot of the camera (5); each fish tank (101) has a sensor group on its side wall and a water replenishment pool (303) and its water replenishment pipeline (102) on its top. The side wall of the fish tank (101) is connected to the feeding mechanism (4) to achieve precise feeding. The electric ball valve (201) in the fish collection layer (2) is connected to the bottom drain of the fish tank (101) through a sealing flange at its input end. The electric ball valve (201) is fixed by a clamp at its output end to an adjustable angle PVC pipe (202). The lower end of the PVC pipe (202) hangs down towards the net-type fish cart (203). The trapdoor (2031) at the rear of the fish cart (203) is connected to the body of the fish cart (203) through a spring hinge. The trapdoor (2031) is locked on both sides by a pin-type snap fastener (2032), forming a gravity unloading mechanism. The water storage layer (3) receives the water leaking from the fish collection layer (2). The three-stage filtration device (301) is installed to fix the PP cotton layer, activated carbon layer and ultrafiltration membrane layer from top to bottom through the chute. It is used to remove impurities, adsorb organic matter and purify microorganisms in stages. After filtration, the water flows into the water storage tank (302) through the guide channel and is pumped into the water replenishment tank (303) of the aquaculture layer (1) by the water pump through the pipeline to form a closed water cycle.

2. The intelligent aquaculture system based on the device described in claim 1, characterized in that: include: The system includes an environmental control module, a feeding control module, a health monitoring module, a water circulation control module, and a central control module. The central control module is connected via electrical lines to a sensor group fixed to the side wall of the aquarium (101) in the breeding layer (1), a camera (5) aimed at the biological behavior observation channel, a servo motor (403) of the top feeding mechanism (4), an electric ball valve (201) at the bottom drain of the aquarium (101), and a water pump in the water storage layer (3). The central control module generates control commands based on the water quality data collected by the sensor group and the fish behavior data captured by the camera (5) to drive the start and stop actions of the servo motor (403), the electric ball valve (201), and the water pump.

3. The intelligent aquaculture system as described in claim 2, characterized in that: The environmental control module reads data from the sensor group installed on the side walls of the two polycarbonate fish tanks (101) in the aquaculture layer (1) in real time, including water temperature, pH value, dissolved oxygen, and ammonia nitrogen concentration; it adjusts the water inflow of the two fish tanks (101) respectively through the independent electric control valve connected to the top flange of the compartmentalized water replenishment tank (303); when the ammonia nitrogen concentration of a certain fish tank (101) exceeds the set threshold of 1.5mg, it increases the opening of the electric control valve corresponding to that fish tank (101) to accelerate water renewal.

4. The intelligent aquaculture system as described in claim 2, characterized in that: The feeding control module controls the rotation angle of the MG996R servo motor (403) in the feeding mechanism (4) that is snapped onto the top of the aquaculture layer (1); the feeding frequency is calculated based on the fish distribution density captured by the camera (5) in the biological behavior observation channel. When the percentage of pixels in the observation channel area where the fish gather is greater than 70%, the servo motor (403) is triggered to open the gate at 10-second intervals. When the percentage of pixels when the fish are dispersed is less than 30%, stop feeding.

5. The intelligent aquaculture system as described in claim 2, characterized in that: The health monitoring module captures the video stream of the fish swimming at 30fps through a camera (5) with a 50mm wide observation channel facing it; it analyzes the frequency of the fish rubbing against the tank wall and the duration of surfacing: if the rubbing frequency is >3 times / minute within 5 minutes, it is marked as abnormal behavior; if the surfacing duration is >2 minutes, it sends a first-level warning signal to the central control module; when the abnormality lasts for 5 minutes, it opens the electric ball valve (201) at the bottom of the corresponding fish tank (101), so that the fish fall into the stainless steel woven mesh of the leaky fish cart (203) through the adjustable angle PVC pipe (202).

6. The intelligent aquaculture system as described in claim 2, characterized in that: The water circulation control module monitors the operating status of the three-stage filtration device (301) in the water storage layer (3), including: detecting turbidity changes through the water quality sensor at the outlet of the guide channel; determining that the ultrafiltration membrane is blocked when the turbidity value increases by 40% compared to the initial value; and controlling the water pump to pump the filtered water from the water storage tank (302) into the replenishment tank (303) of the aquaculture layer (1) through the pipeline to maintain a closed water circulation.

7. The intelligent aquaculture system as described in claim 2, characterized in that: It also includes a fish disease detection module, including: A MaixCam camera (5) is deployed in the biological behavior observation channel of the aquaculture layer (1) to capture real-time video streams of fish schools; The image processing unit, which is connected to the camera (5), has a built-in visual model trained based on the YOLOv5 algorithm to identify fish disease characteristics in the video stream; The server receives disease detection result data output by the image processing unit; The client-side interactive interface is used to display the real-time image of the camera (5) and mark the identified disease type and confidence level, while generating a historical record containing timestamps, disease type and severity. The visual model performs the following: real-time analysis of video frames captured by the camera (5); identification of diseased fish and output of disease type and confidence level; the client interface triggers a system warning when the confidence level exceeds the threshold and supports querying historical records to generate disease statistical reports.

8. The intelligent aquaculture system as described in claim 2, characterized in that: The feeding control module dynamically adjusts the feeding amount based on the weight of the residual feed collected at the bottom of the fish collection layer (2) and the net-type fish cart (203): After the daily feeding is completed, the weight of the residual feed recorded by the weighing sensor is read; if the weight of the residual feed accounts for more than 15% of the total feeding amount of the day, the basic feeding amount of the next day is reduced proportionally; if the weight of the residual feed is less than 5%, the feeding amount of the next day is increased by 10%.

9. The intelligent aquaculture system as described in claim 2, characterized in that: The central control module processes dissolved oxygen data from the sensor group: when the dissolved oxygen concentration is <4mg / L, it immediately starts the oxygenation equipment and opens the water supply valve; when the ammonia nitrogen concentration is >1.5mg / L, it accelerates the operation frequency of the water pump to 120% of the rated power to enhance the water circulation filtration efficiency.

10. The intelligent aquaculture system as described in claim 2, characterized in that: The two independent chambers of the compartmentalized water supply tank (303) are equipped with water level sensors at their bottoms; the water level data of each chamber independently controls the corresponding water supply valve of the aquarium (101): When the water level is 20mm below the calibrated height, the electrically controlled valve will be opened to replenish water. When the water level reaches the set height, the electric control valve is closed and a water replenishment completion signal is sent.

Images