A fish and vegetable symbiosis remote management method, device and equipment based on digital twinning

Abstract

The application discloses a kind of based on digital twinning's fish and plant symbiosis remote management method, device and equipment, comprising: obtaining the physical parameter data of physical fish pond and physical plant shed in fish and plant symbiosis system, and based on physical parameter data creates the fish group, fish pond, hydroponic plant, plant shed and the corresponding twin body model of physical equipment of fish and plant symbiosis system;Real-time acquisition is deployed in fish and plant symbiosis system The fish group breeding parameter, plant growth parameter collected by sensor and camera, and obtain equipment operating parameter;According to fish group breeding parameter, plant growth parameter and equipment operating parameter, analysis is carried out by fish and plant symbiosis digital twinning platform, and analysis result is obtained;According to analysis result, trigger early warning mechanism and generate control instruction, according to control instruction dynamically adjust the operating state of corresponding physical equipment in fish and plant symbiosis system, realize remote collaborative management.Can realize the remote management of fish and plant symbiosis system, improve resource utilization and convenience.

Description

Technical Field

[0001] The present invention relates to the field of digital twin technology, and in particular to a remote management method, device and equipment for fish and vegetable symbiosis based on digital twin. Background Art

[0002] Aquaponics is an ecological agricultural model that combines aquaculture with soilless cultivation. Fish excrement provides nutrients to plants, which then absorb and purify the water before returning it to the pond. This creates a closed-loop ecosystem: "fish fertilize the water, vegetables purify the water, and the water feeds the fish." This system offers significant advantages in resource recycling. However, traditional aquaponics systems rely heavily on manual management, which is labor-intensive and inefficient. Precise control and remote monitoring are also difficult to achieve, making system instability easily caused by human negligence or lack of experience. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to propose a remote management method, device and equipment for fish-vegetable symbiosis based on digital twins, aiming to solve the problems of low efficiency and easy errors in manual management of traditional systems, and to achieve precise linkage control, remote monitoring and intelligent optimization.

[0004] To achieve the above objectives, the present invention provides a remote management method for aquaponics based on digital twins, the method comprising:

[0005] Obtain physical parameter data of the physical fish pond and physical plant shed in the aquaponics system, and create twin models corresponding to the fish school, fish pond, hydroponic plants, plant shed, and physical equipment of the aquaponics system based on the physical parameter data to obtain an aquaponics digital twin platform;

[0006] Real-time acquisition of fish farming parameters, plant growth parameters, and equipment operating parameters collected by sensors and cameras deployed in the aquaponics system, and uploading them to a cloud server;

[0007] Analyzing the fish farming parameters, plant growth parameters, and equipment operation parameters through the aquaponics digital twin platform to obtain analysis results;

[0008] According to the analysis results, an early warning mechanism is triggered and a control instruction is generated. The operating status of the corresponding physical equipment in the fish-vegetable symbiotic system is dynamically adjusted according to the control instruction, thereby realizing remote collaborative management of fish farming, plant growth and equipment operation.

[0009] Preferably, the physical parameter data includes:

[0010] The structural layout parameters of the physical fish pond, including the type of fish pond, layered design, size and shape of the fish pond, and the location of the fish pond;

[0011] The layout parameters of the physical plant shed, including the type of planting racks, the size and shape of the planting racks, and the location of the planting racks;

[0012] Device information of physical devices, including device type, model, serial number, installation location, and device operating parameters.

[0013] Preferably, the step of creating twin models corresponding to the fish school, fish pond, hydroponic plants, plant shed, and physical equipment of the aquaponics system based on the physical parameter data includes:

[0014] A three-dimensional model is constructed based on the structural layout parameters of the physical fish pond, and the environmental parameters of the fish pond are mapped to obtain a twin model of the fish pond including the fish school and the fish pond;

[0015] A three-dimensional model is constructed based on the layout parameters of the physical plant shed, and the plant environment parameters are mapped to obtain a twin model of the plant shed including hydroponic plants and planting racks.

[0016] Create virtual entities for physical equipment including sensors, cameras, water pumps, aerators, ultraviolet sterilizers, microfiltration machines, fill lights, and nutrient solution mixing devices, and synchronize the device information of the physical equipment to obtain a device twin model.

[0017] Preferably, the fish breeding parameters include fish pond water temperature, dissolved oxygen concentration, water quality parameters, pH value, and fish image data; the plant growth parameters include plant shed temperature, humidity, light intensity, carbon dioxide concentration, nutrient solution pH value, and plant image data; the equipment operation parameters include water pump flow, aerator power, ultraviolet sterilizer status, microfiltration machine working cycle, fill light spectrum parameters, ventilation window opening and closing angle, and nutrient solution mixing ratio.

[0018] Preferably, the analysis performed by the aquaponics digital twin platform based on the fish farming parameters, plant growth parameters, and equipment operation parameters includes:

[0019] When the fish pond environmental parameters meet the preset growth environment range, the fish health status of the fish in different fish ponds is analyzed and the growth trend of the fish is predicted based on the fish image data combined with the feeding amount, wherein the fish pond environmental parameters include fish pond water temperature, dissolved oxygen concentration, water quality parameters, and pH value;

[0020] When the plant environmental parameters meet the preset growth environment range, the plant growth status of hydroponic plants in different planting racks is analyzed and the plant growth trend and yield are predicted based on the plant image data combined with the biological characteristics of the plants. The plant environmental parameters include plant shed temperature, humidity, light intensity, carbon dioxide concentration, and nutrient solution pH value.

[0021] When the fish population health status or the plant growth status is abnormal, a historical disease database is called to match and obtain a treatment scheme;

[0022] The device operation parameters are analyzed to detect whether the physical device has a device anomaly, obtain an anomaly processing scheme, and the treatment scheme and / or the anomaly processing scheme are taken as the analysis result.

[0023] Preferably, the warning mechanism triggered according to the analysis result comprises:

[0024] The target abnormal area is marked by a red floating window in the fish-plant symbiosis digital twin platform, and the corresponding fish population, hydroponic plants and / or physical device corresponding to the twin model are displayed in red holographic images.

[0025] Preferably, the running state of the corresponding physical device in the fish-plant symbiosis system is dynamically adjusted according to the regulation instruction, comprising:

[0026] When the analysis result indicates that the dissolved oxygen concentration needs to be adjusted, the oxygen increasing machine is started and the power is adjusted;

[0027] When the analysis result indicates that the water quality parameter needs to be adjusted, the microfilter is started for circulating filtration, the ultraviolet sterilization machine is turned on, and the water pump is started and the flow is adjusted;

[0028] When the analysis result indicates that the light needs to be adjusted, the intensity and spectrum of the light supplementing lamp are adjusted to the wave band required by the plants;

[0029] When the analysis result indicates that the temperature needs to be adjusted, the opening angle of the ventilation window is adjusted;

[0030] When the analysis result indicates that the carbon dioxide concentration needs to be adjusted, the ventilation device or the supplementary air source is started;

[0031] When the analysis result indicates that the pH value of the nutrient solution needs to be adjusted, the nutrient solution dispensing device is started and the ratio is adjusted.

[0032] To achieve the above purpose, the application also provides a fish-plant symbiosis remote management device based on digital twinning, which comprises:

[0033] A model construction unit is configured to acquire physical parameter data of a physical fish tank and a physical plant shed in a fish-plant symbiosis system, and create fish population, fish tank, hydroponic plant, plant shed and physical device corresponding to the twin model of the fish-plant symbiosis system based on the physical parameter data, to obtain a fish-plant symbiosis digital twin platform;

[0034] A data acquisition unit is configured to acquire fish population breeding parameters and plant growth parameters collected by sensors and cameras deployed in the fish-plant symbiosis system in real time, and acquire device operation parameters and upload them to a cloud server.

[0035] a data analysis unit, configured to analyze the fish farming parameters, plant growth parameters, and equipment operation parameters through the aquaponics digital twin platform to obtain analysis results;

[0036] The remote management unit is used to trigger the early warning mechanism and generate control instructions based on the analysis results, and dynamically adjust the operating status of the corresponding physical equipment in the fish-vegetable symbiotic system according to the control instructions to achieve remote collaborative management of fish farming, plant growth and equipment operation.

[0037] In order to achieve the above-mentioned objectives, the present invention also proposes a remote management device for fish and vegetable symbiosis based on digital twins, comprising a processor, a memory, and a computer program stored in the memory, wherein the computer program is executed by the processor to implement the steps of a remote management method for fish and vegetable symbiosis based on digital twins as described in the above-mentioned embodiment.

[0038] In order to achieve the above-mentioned objectives, the present invention also proposes a computer-readable storage medium, on which a computer program is stored. The computer program is executed by a processor to implement the steps of a remote management method of fish-vegetable symbiosis based on digital twins as described in the above-mentioned embodiment.

[0039] In order to achieve the above-mentioned objectives, the present invention also proposes a computer program product, including a computer program / instruction, which, when executed by a processor, implements the steps of a remote management method of fish and vegetable symbiosis based on digital twins as described in the above-mentioned embodiment.

[0040] Beneficial effects:

[0041] The above solution, by building a digital twin platform for fish-vegetable symbiosis to integrate fish farming, plant growth and equipment operation data, can achieve remote collaborative management of the fish-vegetable symbiosis system, dynamically adjust the equipment operation status through analysis results, ensure stable system operation, and reduce losses caused by abnormal parameters. In addition, based on the analysis and regulation of real-time data, it can more accurately meet the growth needs of fish and plants and avoid waste of resources, thereby improving resource utilization efficiency, management efficiency and convenience, and reducing manual inspection costs. It is especially suitable for large-scale farms or unmanned scenarios.

[0042] Clarifying the detailed parameters of the physical fish pond, plant shed, and equipment provides the foundation for building an accurate twin model. This allows the digital twin platform to more realistically reflect the state and characteristics of the physical system, thereby improving management accuracy. Based on this parameter information, a three-dimensional model is constructed, environmental parameters are mapped, and virtual entities are created for the physical equipment. This allows for precise simulation of each component of the aquaponics system, providing detailed virtual model support for subsequent analysis and control, enhancing the credibility of the virtual simulation, and ensuring dynamic consistency between the virtual platform and the physical system.

[0043] It covers a full range of parameters related to fish farming, plant growth, and equipment operation, enabling comprehensive monitoring of the operating status of the fish-vegetable symbiotic system. This provides a rich data basis for subsequent precise analysis and regulation, and can promptly capture various changes and problems in the system. Moreover, different types of parameter data can reflect the operating status of the system from multiple dimensions, such as the health status of fish, the growth trend of plants, the operating efficiency of equipment, etc., which helps to conduct comprehensive analysis and evaluation and formulate more reasonable management strategies.

[0044] By analyzing the environmental parameters of fish ponds and plant sheds, combined with image data and biological characteristics, it is possible to accurately analyze the health status and growth trends of fish schools, as well as the growth status and yield forecasts of plants, providing a strong basis for scientific management and making production plans and resource allocations in advance; when there are growth abnormalities or equipment abnormalities, it is possible to call the historical disease database or perform equipment abnormality detection to obtain corresponding treatment plans and abnormality handling plans, and take timely measures to solve problems, reduce losses, and ensure the normal operation of the system.

[0045] The target abnormal area is marked with a red floating window, and the corresponding twin model is displayed as a red holographic image, which can intuitively show the user the abnormal situation in the system, allowing the user to quickly locate the problem, improve the timeliness and efficiency of problem handling, and facilitate overall management and decision-making.

[0046] Based on the analysis results, control instructions are automatically generated and the operating status of physical equipment is dynamically adjusted, realizing automated control of equipment, reducing manual intervention, improving the accuracy and efficiency of management, and being able to quickly respond to various changes and needs in the system; then, specific control measures are taken for different abnormal environmental parameters, such as adjusting dissolved oxygen concentration, water quality parameters, light, temperature, carbon dioxide concentration and nutrient solution pH value, etc., which can finely control the growth environment of the fish-vegetable symbiotic system, provide more suitable growth conditions for fish and plants, and promote their healthy growth and high yield. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0048] Figure 1 A flowchart of a remote management method for aquaponics based on digital twins is provided in accordance with an embodiment of the present invention.

[0049] Figure 2 This is a display effect diagram of a fish growth laboratory provided by one embodiment of the present invention.

[0050] Figure 3 This is a display effect diagram of a plant growth laboratory provided by one embodiment of the present invention.

[0051] Figure 4 A schematic structural diagram of a remote management device for aquaponics based on digital twins provided in one embodiment of the present invention.

[0052] The realization of the objectives of the invention, the functional features and advantages will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the invention for which protection is sought, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0054] The present invention is described in detail below with reference to the embodiments.

[0055] Reference Figure 1 The figure shows a flow chart of a remote management method of aquaponics based on digital twins provided by one embodiment of the present invention.

[0056] In this embodiment, the method includes:

[0057] S11, obtain the physical parameter data of the physical fish pond and physical plant shed in the aquaponics system, and create twin models corresponding to the fish school, fish pond, hydroponic plants, plant shed and physical equipment of the aquaponics system based on the physical parameter data to obtain the aquaponics digital twin platform.

[0058] Furthermore, the physical parameter data includes:

[0059] The structural layout parameters of the physical fish pond, including the type of fish pond, layered design, size and shape of the fish pond, and the location of the fish pond;

[0060] The layout parameters of the physical plant shed, including the type of planting racks, the size and shape of the planting racks, and the location of the planting racks;

[0061] Device information of physical devices, including device type, model, serial number, installation location, and device operating parameters.

[0062] Furthermore, in step S11, the twin models corresponding to the fish school, fish pond, hydroponic plants, plant shed and physical equipment of the aquaponics system are created based on the physical parameter data, including:

[0063] S11-1, constructing a three-dimensional model based on the structural layout parameters of the physical fish pond, and mapping the environmental parameters of the fish pond to obtain a twin model of the fish pond including the fish school and the fish pond;

[0064] S11-2, constructing a three-dimensional model based on the layout parameters of the physical plant shed, and mapping the plant environment parameters to obtain a twin model of the plant shed including hydroponic plants and planting racks;

[0065] S11-3, establish virtual entities for physical equipment including sensors, cameras, water pumps, aerators, ultraviolet sterilizers, microfiltration machines, fill lights and nutrient solution mixing devices, and synchronize the device information of the physical equipment to obtain a device twin model.

[0066] In this embodiment, based on the fish, breeding equipment and environmental parameters of the fish-vegetable symbiotic system, as well as the hydroponic plants, planting equipment and environmental parameters, a corresponding twin model is constructed based on the Python framework according to the proportion of the fish-vegetable symbiotic system, and the physical equipment and the equipment twin model are matched by setting sequence numbers to build a fish-vegetable symbiotic digital twin platform. Figure 2 and Figure 3As shown, the fish-plant symbiosis digital twin platform includes a fish-plant symbiosis laboratory, which is divided into two floors. The first floor is a fish population growth laboratory, and the second floor is a hydroponic plant growth laboratory. The first floor fish population growth laboratory includes different types of fish pools (including fish breeding pools, sedimentation pools, square pools, water storage pools, biochemical pools), water pumps, control cabinets, micro-filtration machines, oxygenators, ultraviolet sterilizers, sensors (including water temperature sensors, dissolved oxygen sensors, water quality detectors, humidity sensors, pressure sensors), and intelligent cameras. The second floor hydroponic plant growth laboratory includes different types of planting racks (including T-shaped cultivation racks and square cultivation racks), environmental monitors, plant light supplement lamps, sensors (including pH sensors, light sensors, and humidity sensors), and intelligent cameras.

[0067] By determining the setting positions and setting quantities of different twin models in the fish-plant symbiosis digital twin platform according to the relevant position information in the fish-plant symbiosis system, the fish-plant symbiosis digital twin platform can accurately simulate the setting details of the physical devices of the real fish-plant symbiosis system, further improving the accuracy and real-time performance of monitoring. Further, a device file is established for each physical device to record the type, model, specification, number, and other information of the physical device, so as to facilitate quick management and query. In addition, according to the correspondence between fish tags and different entity fish, the GIS (Geographic Information System) of the fish pool is used to determine the corresponding fish pool and the current position and motion trajectory of the virtual fish and the entity fish. Different sensors are arranged on each planting rack of the plant shed in the fish-plant symbiosis system, and each sensor is used as an identifier of the planting rack to facilitate the recording of the growth of plants on the planting rack, such as three-color lights and intelligent cameras arranged on each planting rack.

[0068] S12, real-time acquisition of fish population breeding parameters, plant growth parameters collected by sensors and cameras deployed in the fish-plant symbiosis system, and acquisition of device operation parameters, and uploading to a cloud server.

[0069] Further, the fish population breeding parameters include fish pool water temperature, dissolved oxygen concentration, water quality parameters, pH value, and fish image data. The plant growth parameters include plant shed temperature, humidity, light intensity, carbon dioxide concentration, nutrient solution pH value, and plant image data. The device operation parameters include water pump flow, oxygenator power, ultraviolet sterilizer state, micro-filtration machine working period, light supplement lamp spectrum parameters, ventilation window opening angle, and nutrient solution blending ratio.

[0070] S13, analysis by the fish-plant symbiosis digital twin platform according to the fish population breeding parameters, plant growth parameters, and device operation parameters to obtain an analysis result.

[0071] Further, in step S13, the fishery parameters, plant growth parameters, and equipment operation parameters are analyzed by the fish-plant symbiotic digital twin platform, including:

[0072] S13-1, when the fish tank environment parameters meet the preset growth environment range, the fish group health state analysis and fish group growth trend prediction of different fish tanks are performed according to the fish image data combined with the feeding amount, wherein the fish tank environment parameters include fish tank water temperature, dissolved oxygen concentration, water quality parameters, and pH value;

[0073] S13-2, when the plant environment parameters meet the preset growth environment range, the plant growth state analysis and plant growth trend and yield prediction of different planting racks of water plants are performed according to the plant image data combined with the biological characteristics of the plants, wherein the plant environment parameters include plant shed temperature, humidity, light intensity, carbon dioxide concentration, and nutrient solution pH value;

[0074] S13-3, when the fish group health state or the plant growth state is abnormal, the historical disease database is called to match to obtain a treatment scheme;

[0075] S13-4, the equipment operation parameters are analyzed to detect whether the physical equipment has equipment abnormalities, obtain an abnormal treatment scheme, and the treatment scheme and / or the abnormal treatment scheme are taken as the analysis result.

[0076] Wherein, the fish group health state analysis and fish group growth trend prediction of different fish tanks are performed according to the fish image data combined with the feeding amount, including:

[0077] The fish image data is preprocessed and image feature extraction is performed to obtain fish apparent characteristics;

[0078] The fish apparent characteristics are input into the constructed health state analysis model and fish growth prediction model respectively to obtain health state analysis results and fish growth prediction results respectively.

[0079] Specifically, the fish images of the fish group are collected in real time by the intelligent camera deployed in the fish tank, and the fish images include different fish species, postures and growth stages; the fish images are labeled, including labeling the fish species, body length, body surface lesions (such as white spots and ulcers), gill state and other health indicators, and recording the corresponding feeding amount data; the fish images are preprocessed including data enhancement fish standardization; the pre-trained ResNet50 model (without the top classification layer) is used to extract image features, and the output of the residual block is reserved as the fish apparent feature; further, the feature pyramid network (FPN) can be used to integrate shallow details (such as lesion edges) and deep semantic information (such as overall morphology) for multi-scale feature fusion. The health state analysis model based on the double-branch multi-task learning model (DB-MTL) is constructed, the fish apparent feature is input into the health state analysis model, and the health state analysis results including the health level (such as healthy, sub-healthy and diseased) and the disease type (such as parasitic and bacterial infection) are output.

[0080] The fish apparent feature, the fish group environment parameter and the historical time sequence (such as the growth rate for 7 days, the body length / weight change) are input into the fish growth prediction model based on the LSTM model, and the fish growth prediction results of the body weight growth rate and the body length change for the next 7 days are output. The architecture of the model includes an input layer (multi-modal features after fusion of the fish apparent feature, the fish group environment parameter and the historical time sequence), an LSTM layer (including 128 hidden units to capture time dependence) and an output layer for prediction, and the mean square error (MSE) loss function is used to optimize the growth trend prediction accuracy.

[0081] The plant growth state of the water-grown plants in different planting racks is analyzed and the plant growth trend and yield are predicted according to the plant image data combined with the biological characteristics of the plants, including:

[0082] The plant image data is subjected to image segmentation and image feature extraction to obtain multi-spectral features;

[0083] The multi-spectral features are input into the constructed growth state analysis model and the plant growth prediction model respectively to obtain the growth state analysis results and the plant growth prediction results.

[0084] Specifically, plant images (including leaf images) are collected by hyperspectral cameras, covering visible light (400-700 nm) and near-infrared (700-1000 nm) bands; key growth indicators are labeled on the plant images, including leaf area index (LAI), chlorophyll content, and disease spot area (yellowing, brown spots); the U-Net model is used to segment the leaf area of the plant image, remove background noise, and extract multispectral features, including texture features (such as gray level co-occurrence matrix (GLCM) calculation of contrast, energy, and homogeneity), morphological features (such as disease spot area ratio, edge irregularity (fractal dimension)), and spectral features (such as near-infrared band reflectance (negatively correlated with chlorophyll content)). The multispectral features and plant environmental parameters (such as CO2 concentration, light intensity, and nutrient solution pH) are input into the growth state analysis model (including random forest classifier and isolated forest algorithm), and the growth state analysis results including growth state level (such as normal, mild stress, and severe stress) and plant disease type (identification of abnormal samples deviating from normal distribution (such as nutrient deficiency)) are output.

[0085] The seasonal cycles (such as day and night, seasonal changes) of light and temperature data are decomposed using the SARIMA model to predict the trends of environmental parameters in the next 30 days; the multispectral features and environmental parameter trends are input into the plant growth prediction model based on the LSTM-Transformer hybrid model, and the plant growth prediction results of LAI and yield in the next 30 days are output. The model architecture includes input layer (multispectral features, historical LAI sequence, and SARIMA predicted environmental parameter trends), LSTM layer (capturing long-term dependencies), Transformer encoder (weighted attention to key growth stages (such as flowering period)), and output layer for prediction. The model parameters are updated based on real-time image data and environmental parameters, and the sliding window mechanism (such as window size = 30 days) is used to continuously optimize the prediction accuracy. If fish diseases cause water quality deterioration (such as increased ammonia nitrogen), by analyzing plant root absorption efficiency, it can predict the associated growth abnormalities, and realize closed-loop optimization of the fish-vegetable symbiotic system.

[0086] Furthermore, in another embodiment, by establishing a mathematical model including a water quality model, a plant growth model, and a fish growth model as historical data, the fish-vegetable symbiotic digital twin platform is used to analyze the sensor data and image data collected by sensors and cameras at different locations in real time based on the historical data, and the analysis results are saved and the historical data is updated. Fish school status is analyzed and predicted based on fish images and food intake. For example, the system determines whether the current food intake of a fish school is declining and whether the fish's growth environment parameters are within the growth range. Environmental parameters include water quality (including ammonia nitrogen, nitrite, hydrogen sulfide, total alkalinity, and total hardness), temperature, humidity, and various air gases (including ammonia, carbon dioxide, sulfur dioxide, and nitrogen). Specifically, the system determines whether the currently collected values ​​are within the optimal water temperature, humidity, water quality, and air gas thresholds. If they are outside these ranges, a warning is issued. If they are within these ranges, the system analyzes and predicts the fish school status based on the fish images and food intake. If the current state of the fish school shows abnormal growth, the system uses a historical disease database for matching and identifying similar treatment plans. The system also predicts the growth of the fish school by extracting features from fish images (such as color abnormalities, texture changes, and morphological characteristics), detecting surface lesions (ulcers, white spots, and congestion), and identifying tissue changes invisible to the naked eye (such as changes in reflectivity caused by gill hypoxia).

[0087] Similarly, plant image data combined with biological characteristics can be used to analyze the growth status of hydroponic plants and predict their growth trends and yield. By collecting plant images and performing feature extraction, we can capture surface characteristics of leaves (e.g., yellowing and browning) such as vegetable leaves, as well as local textures of spots, mold, and holes (using gray-level co-occurrence matrix (GLCM) analysis), the edge contours (circular, irregular) and distribution density of lesions, and reflectance anomalies in specific bands of multispectral data (e.g., near-infrared bands are sensitive to chlorophyll) to analyze plant growth. Based on fish images and biological characteristics of plant growth, such as variety, growth stage, and physiological indicators, we can predict the growth trend and yield of hydroponic plants. For example, we use the SARIMA (Seasonal Autoregressive Integrated Moving Average) algorithm to account for the influence of periodic environmental factors such as light and temperature, and the LSTM (Long Short-Term Memory) algorithm to capture nonlinear growth dynamics. By inputting environmental parameters (CO2 concentration, temperature and humidity) and a time series of leaf area index (LAI), we ultimately output predicted stem height and leaf area values ​​for the next seven days. It also links whether abnormal fish health will have corresponding abnormal effects on plant growth, and conversely whether plant diseases will have an impact on fish growth, and records these data in the digital twin platform for data improvement.

[0088] The water quality model collects water-related data and information through various sensors (such as pressure sensors, flow sensors, and water quality sensors) deployed at key locations such as fish ponds, square ponds, reservoirs, biochemical ponds, and water pumps. These sensors can sense and measure various physical and chemical parameters in the environment or water, such as pH value, dissolved oxygen, turbidity, and temperature, in real time. The collected sensor data can be converted into electrical signals and subjected to preliminary processing and conversion, usually including operations such as signal amplification, filtering, and analog-to-digital conversion, to convert them into digital signals for subsequent analysis. Data transmission technology transmits the processed sensor data to the cloud server via wired or wireless means (such as based on LoRa wireless communication technology).

[0089] The establishment of a plant growth model involves simulating the growth of hydroponic plants under different environmental conditions based on the biological principles of plant growth and by collecting monitoring data from the hydroponic plant growth process. Specifically, the establishment of a plant growth model involves using deployed sensors to collect environmental parameters such as temperature, humidity, light, and carbon dioxide concentration in the plant growth environment. (For example, a soil moisture sensor measures the dielectric constant of the soil to accurately determine soil moisture content; a light sensor accurately senses light intensity, duration, and spectral distribution, providing data support for plant photosynthesis; and a carbon dioxide sensor monitors the concentration of carbon dioxide in the air in real time, as carbon dioxide is a key raw material for plant photosynthesis and an appropriate concentration helps improve plant photosynthetic efficiency.)

[0090] Building a fish growth model involves collecting fish image data that covers as many fish species as possible, including diverse environmental scenarios, fish poses and angles, and various growth stages, to ensure accurate recognition in all conditions. This image data can come from databases of scientific research institutions, surveillance footage from fish farming operations, and photos of fish taken by enthusiast photographers. Collected fish images must be accurately labeled, including information such as species, sex (if distinguishing features are available), and age range. In addition to convolutional neural networks (CNNs), deep CNN models such as ResNet50 can also be used for training. ResNet50 introduces residual blocks to address the vanishing gradient problem during deep network training. Each residual block consists of a main path (performing regular convolution operations) and a shortcut path (passing the input directly to the output, where the two are summed to form the final output). This structure facilitates direct information flow and maintains the performance of deep networks. When building a fish recognition model based on ResNet50, a series of operations are required, such as data augmentation (applying techniques such as rotation, flipping, scaling, and color conversion to increase sample diversity and prevent overfitting), removing the fully connected layers of the original model and replacing them with new fully connected layers adapted to the fish recognition task (the number of output nodes is equal to the number of fish species), freezing most of the convolutional layers in the initial training stage and only fine-tuning the newly added fully connected layers to speed up convergence, etc.

[0091] The following further describes the network architecture and training process of the fish recognition model, which addresses the dual tasks of growth stage classification and disease state recognition. The improved dual-branch multi-task learning model (DB-MTL) employs ConvNeXt-L as a shared feature extractor, combined with a dynamic sparse attention (DSA) mechanism to enhance sensitivity to fish texture. The growth stage branch adds non-local blocks to capture global fish morphological features (such as body length and fatness). The disease state recognition branch uses a YOLOv8-P2 architecture to output disease type and lesion location (Bounding Box). A gradient reversal layer (GRL) separates growth-related features from disease features to avoid cross-task interference. A BiFPN architecture is inserted into the backbone network to enhance detection of small lesions (such as white spot disease). Depthwise separable convolutions replace some standard convolutions, reducing the model size.

[0092] S14, triggering an early warning mechanism and generating a control instruction based on the analysis results, and dynamically adjusting the operating status of the corresponding physical equipment in the fish-vegetable symbiotic system according to the control instruction to achieve remote collaborative management of fish farming, plant growth and equipment operation.

[0093] Furthermore, in step S14, triggering the early warning mechanism according to the analysis results includes:

[0094] The target abnormal area is marked with a red floating window in the fish-vegetable symbiotic digital twin platform, and the twin models of the corresponding fish schools, hydroponic plants and / or physical equipment are displayed as red holographic images.

[0095] Furthermore, in step S14, dynamically adjusting the operating state of corresponding physical devices in the aquaponics system according to the control instructions includes:

[0096] S14-1, when the analysis result indicates that the dissolved oxygen concentration needs to be adjusted, starting the aerator and adjusting the power;

[0097] S14-2, when the analysis result indicates that the water quality parameters need to be adjusted, the microfiltration machine is started for circulating filtration, the ultraviolet sterilizer is turned on, and the water pump is started and the flow rate is adjusted;

[0098] S14-3, when the analysis result indicates that the lighting needs to be adjusted, adjusting the intensity and spectrum of the supplementary light to the wavelength band required by the plants;

[0099] S14-4, when the analysis result indicates that the temperature should be adjusted, adjusting the opening and closing angles of the ventilation windows;

[0100] S14-5, when the analysis result indicates that the carbon dioxide concentration should be adjusted, starting ventilation equipment or supplementing the air source;

[0101] S14-6, when the analysis result indicates that the pH value of the nutrient solution should be adjusted, the nutrient solution mixing device is started and the mixing ratio is adjusted.

[0102] In this embodiment, based on the analysis results obtained above, early warnings are triggered and control instructions are generated to remotely control the operating status of physical equipment. This is based on a PLC controller, which generates control instructions based on the output of the digital twin platform, dynamically adjusting the operating status of physical equipment, including water pumps, aerators, LED lighting equipment, temperature control equipment, and nutrient solution mixing devices. If the analysis results indicate that oxygen should be added to the fish pond, an oxygenation command is generated and sent to the aerator of the physical equipment. If the data received indicates that there is sufficient oxygen and no further oxygen is needed, the digital twin platform generates a command to stop the growth and sends it to the physical equipment. Similarly, if the vegetables are not getting enough light, a command to turn on the plant fill light is issued and sent to the physical equipment fill light. If the data received indicates that no more light is needed, a command to stop the plant fill light is issued and sent to the physical equipment fill light.

[0103] At the same time, users can remotely control physical devices based on their needs through the corresponding twin models on the digital twin platform. Triggering an alert marks the target abnormal area with a red floating window and displays the corresponding twin models of fish, hydroponic plants, and / or physical devices as red holographic images.

[0104] For example, if the analysis results indicate that the temperature of a corresponding vegetable rack is too high, a red floating window will be displayed on the rack as a warning, and a control instruction will be generated to adjust the operating status of the corresponding physical device. A preset step interface for the current operation is displayed, including whether to open the ventilation window, allowing the user to manually select the operation option for the corresponding physical device, thereby issuing an opening command for all corresponding ventilation windows. If the corresponding vegetable rack is insufficiently illuminated, a red floating window will appear on the rack as a warning, and a control instruction will be generated to adjust the operating status of the corresponding physical device. A preset step interface for the current operation is displayed, including whether to turn on the plant fill light, allowing the user to manually select the operation option for the corresponding physical device, thereby issuing an opening command for the corresponding plant fill light. If the fish pond is insufficiently oxygenated, a red floating window will appear on the pond as a warning, and a control instruction will be generated to adjust the operating status of the corresponding physical device. A preset step interface for the current operation is displayed, including whether to turn on the aerator, allowing the user to manually select the operation option for the corresponding physical device, thereby issuing an opening command for the corresponding fish pond aerator. When the water quality monitoring of the fish pond indicates abnormality, a red floating window warning prompt will appear in the fish pond, and a control instruction will be generated to adjust the operating status of the corresponding physical equipment. The preset step interface that needs to be operated currently can also be displayed, including whether to turn on ultraviolet sterilization, whether to turn on the microfiltration machine, and whether to turn on the circulating water pump, so that the user can manually check the operation options of the corresponding physical equipment to issue a start command to the corresponding fish pond aerator ultraviolet sterilizer and circulating water pump.

[0105] Reference Figure 4 Shown is a structural schematic diagram of a remote management device for aquaponics based on digital twins provided by one embodiment of the present invention.

[0106] In this embodiment, the device 20 includes:

[0107] A model building unit 21 is used to obtain physical parameter data of the physical fish pond and physical plant shed in the aquaponics system, and create twin models corresponding to the fish school, fish pond, hydroponic plants, plant shed and physical equipment of the aquaponics system based on the physical parameter data to obtain an aquaponics digital twin platform;

[0108] The data acquisition unit 22 is used to obtain fish farming parameters, plant growth parameters, and equipment operation parameters collected by sensors and cameras deployed in the aquaponics system in real time, and upload them to the cloud server;

[0109] a data analysis unit 23 for analyzing the fish farming parameters, plant growth parameters, and equipment operation parameters through the aquaponics digital twin platform to obtain analysis results;

[0110] The remote management unit 24 is used to trigger the early warning mechanism and generate control instructions based on the analysis results, and dynamically adjust the operating status of the corresponding physical equipment in the fish-vegetable symbiotic system according to the control instructions to achieve remote collaborative management of fish farming, plant growth and equipment operation.

[0111] Each unit module of the device 20 can respectively execute the corresponding steps in the above method embodiment, so each unit module will not be described in detail here. Please refer to the description of the corresponding steps above for details.

[0112] The embodiment of the present invention further provides a remote management device for fish and vegetable symbiosis based on digital twins, which includes the remote management device for fish and vegetable symbiosis based on digital twins as described above, wherein the remote management device for fish and vegetable symbiosis based on digital twins can be used. Figure 4 The structure of the embodiment can be executed accordingly. Figure 1 The technical solution of the method embodiment shown has similar implementation principles and technical effects. For details, please refer to the relevant records in the above embodiments and will not be repeated here.

[0113] The device includes: a mobile phone, digital camera, tablet computer, or other device with a camera function, or a device with an image processing function, or a device with an image display function. The device may include components such as a memory, a processor, an input unit, a display unit, and a power supply.

[0114] Among them, the memory can be used to store software programs and modules, and the processor executes various functional applications and data processing by running the software programs and modules stored in the memory. The memory may mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as an image playback function, etc.), etc.; the data storage area can store data created according to the use of the device, etc. In addition, the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage device. Accordingly, the memory may also include a memory controller to provide the processor and the input unit with access to the memory.

[0115] The input unit can be used to receive input digital, character, or image information, and generate keyboard, mouse, joystick, optical, or trackball signal input related to user settings and function control. Specifically, the input unit of this embodiment can include not only a camera, but also a touch-sensitive surface (such as a touch display) and other input devices.

[0116] The display unit can be used to display information input by the user or information provided to the user and various graphical user interfaces of the device, which can be composed of graphics, text, icons, videos and any combination thereof. The display unit may include a display panel. Optionally, the display panel can be configured in the form of an LCD (Liquid Crystal Display), an OLED (Organic Light-Emitting Diode), etc. Furthermore, the touch-sensitive surface can cover the display panel. When the touch-sensitive surface detects a touch operation on or near it, it is transmitted to the processor to determine the type of touch event. The processor then provides a corresponding visual output on the display panel based on the type of touch event.

[0117] The embodiment of the present invention further provides a computer-readable storage medium, which may be a computer-readable storage medium included in the memory in the above embodiment; or a computer-readable storage medium that exists independently and is not assembled into a device. The computer-readable storage medium stores at least one instruction, which is loaded and executed by a processor to implement Figure 1 The computer-readable storage medium may be a read-only memory, a disk, or an optical disk.

[0118] The embodiment of the present invention further provides a computer program product, including a computer program / instruction, which is loaded and executed by a processor to implement Figure 1 A remote management method for aquaponics based on digital twins is shown.

[0119] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. For similar or identical parts between the various embodiments, reference can be made to each other. For the apparatus embodiments, device embodiments, and storage medium embodiments, since they are generally similar to the method embodiments, their descriptions are relatively simple. For relevant parts, reference can be made to the descriptions of the method embodiments.

[0120] Also, as used herein, the terms "comprise", "comprising", or any other variation thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can also include other elements not expressly listed or inherent to such process, method, article, or apparatus. An element proceeded by "comprises... a" does not, without more constraints, exclude the presence of additional identical elements in the process, method, article, or apparatus that comprises the element.

[0121] The foregoing description illustrates and describes the preferred embodiments of the present application only. It will be understood by those skilled in the art that the present application is not limited to the particular embodiments described herein but is capable of a variety of modifications without departing from the scope of the present application. The application includes any alterations and modifications in the art within the scope of the attached claims.

Claims

1. A remote management method for aquaponics based on digital twins, characterized in that: The method comprises: Obtain physical parameter data of the physical fish pond and physical plant shed in the aquaponics system, and create twin models corresponding to the fish school, fish pond, hydroponic plants, plant shed, and physical equipment of the aquaponics system based on the physical parameter data to obtain an aquaponics digital twin platform; Real-time acquisition of fish farming parameters, plant growth parameters, and equipment operating parameters collected by sensors and cameras deployed in the aquaponics system, and uploading them to a cloud server; Analyzing the fish farming parameters, plant growth parameters, and equipment operation parameters through the aquaponics digital twin platform to obtain analysis results; The analysis performed by the aquaponics digital twin platform based on the fish farming parameters, plant growth parameters, and equipment operation parameters includes: When the fish pond environmental parameters meet the preset growth environment range, the fish health status of the fish in different fish ponds is analyzed and the growth trend of the fish is predicted based on the fish image data combined with the feeding amount, wherein the fish pond environmental parameters include fish pond water temperature, dissolved oxygen concentration, water quality parameters, and pH value; When the plant environmental parameters meet the preset growth environment range, the plant growth status of hydroponic plants in different planting racks is analyzed and the plant growth trend and yield are predicted based on the plant image data combined with the biological characteristics of the plants. The plant environmental parameters include plant shed temperature, humidity, light intensity, carbon dioxide concentration, and nutrient solution pH value. When there are abnormalities in the health status of fish or plant growth, the historical disease database is called for matching and a treatment plan is obtained; Analyzing the operating parameters of the device, detecting whether the physical device has an abnormality, obtaining an abnormality treatment plan, and using the treatment plan and / or the abnormality treatment plan as the analysis result; Among them, the fish image data combined with the feeding amount are used to analyze the health status of fish in different fish ponds and predict the growth trend of fish schools, including: Preprocessing the fish image data and extracting image features to obtain the fish appearance characteristics; The fish appearance characteristics are input into the constructed health status analysis model and fish growth prediction model respectively to obtain health status analysis results and fish growth prediction results respectively; The health status analysis model is constructed based on a dual-branch multi-task learning model, which inputs the fish's appearance characteristics into the health status analysis model and outputs health status analysis results including health level and disease type; The fish growth prediction model is constructed based on the LSTM model, which inputs the fish's appearance characteristics, fish environmental parameters and historical time series into the fish growth prediction model and outputs the fish growth prediction results of weight growth rate and body length change in the next 7 days; Among them, the plant growth status of hydroponic plants in different planting racks is analyzed based on plant image data combined with the biological characteristics of the plants, and the plant growth trend and yield are predicted, including: Perform image segmentation and image feature extraction on plant image data to obtain multispectral features; Inputting the multispectral features into the constructed growth state analysis model and plant growth prediction model respectively to obtain growth state analysis results and plant growth prediction results respectively; The growth status analysis model is constructed based on a random forest classifier and an isolation forest algorithm, multispectral features and plant environmental parameters are input into the growth status analysis model, and the growth status analysis results including growth status grade and plant disease type are output; The plant growth prediction model is built based on the LSTM-Transformer hybrid model, which combines multispectral features with environmental parameter trends to input the plant growth prediction model and outputs plant growth prediction results for the next 30 days, including LAI and yield. An early warning mechanism is triggered based on the analysis results and a control instruction is generated. The operating status of the corresponding physical equipment in the aquaponics system is dynamically adjusted according to the control instruction, thereby realizing remote collaborative management of fish farming, plant growth and equipment operation; The method also includes a fish recognition model constructed based on a dual-branch multi-task learning model to achieve the dual-task requirements of growth stage classification and disease state identification. The fish recognition model uses ConvNeXt-L as a shared feature extractor and is combined with a dynamic sparse attention mechanism; wherein, the fish recognition model includes a growth stage classification branch and a disease state identification branch, the growth stage classification branch includes a Non-local Blocks module for capturing the global morphological characteristics of the fish body, the disease state identification branch includes a YOLOv8-P2 structure for outputting the disease type and lesion location, the fish recognition model separates growth-related features from disease features through a gradient inversion layer, embeds a BiFPN structure in the backbone network, and uses depthwise separable convolution for model lightweight processing.

2. The remote management method of aquaponics based on digital twin according to claim 1, characterized in that: The physical parameter data includes: The structural layout parameters of the physical fish pond, including the type of fish pond, layered design, size and shape of the fish pond, and the location of the fish pond; The layout parameters of the physical plant shed, including the type of planting racks, the size and shape of the planting racks, and the location of the planting racks; Device information of physical devices, including device type, model, serial number, installation location, and device operating parameters.

3. The remote management method of aquaponics based on digital twin according to claim 2, characterized in that: The step of creating twin models corresponding to the fish school, fish pond, hydroponic plants, plant shed, and physical equipment of the aquaponics system based on the physical parameter data includes: A three-dimensional model is constructed based on the structural layout parameters of the physical fish pond, and the environmental parameters of the fish pond are mapped to obtain a twin model of the fish pond including the fish school and the fish pond; A three-dimensional model is constructed based on the layout parameters of the physical plant shed, and the plant environment parameters are mapped to obtain a twin model of the plant shed including hydroponic plants and planting racks. Create virtual entities for physical equipment including sensors, cameras, water pumps, aerators, ultraviolet sterilizers, microfiltration machines, fill lights, and nutrient solution mixing devices, and synchronize the device information of the physical equipment to obtain a device twin model.

4. The remote management method of aquaponics based on digital twin according to claim 1, characterized in that: The fish breeding parameters include fish pond water temperature, dissolved oxygen concentration, water quality parameters, pH value, and fish image data; the plant growth parameters include plant shed temperature, humidity, light intensity, carbon dioxide concentration, nutrient solution pH value, and plant image data; the equipment operation parameters include water pump flow, aerator power, ultraviolet sterilizer status, microfiltration machine working cycle, fill light spectrum parameters, ventilation window opening and closing angle, and nutrient solution mixing ratio.

5. The remote management method of aquaponics based on digital twin according to claim 1, characterized in that: The triggering of the early warning mechanism according to the analysis results includes: The target abnormal area is marked with a red floating window in the fish-vegetable symbiotic digital twin platform, and the twin models of the corresponding fish schools, hydroponic plants and / or physical equipment are displayed as red holographic images.

6. The remote management method of aquaponics based on digital twin according to claim 1, characterized in that: The dynamically adjusting the operating state of the corresponding physical equipment in the aquaponics system according to the control instructions includes: When the analysis result indicates that the dissolved oxygen concentration should be adjusted, starting the aerator and adjusting the power; When the analysis result indicates that the water quality parameters need to be adjusted, the microfiltration machine is started for circulating filtration, the ultraviolet sterilizer is turned on, and the water pump is started and the flow rate is adjusted; When the analysis result indicates that the lighting needs to be adjusted, the intensity of the supplementary light and the spectrum are adjusted to the wavelength band required by the plants; When the analysis result indicates that the temperature should be adjusted, adjusting the opening and closing angles of the ventilation windows; When the analysis result indicates adjustment of the carbon dioxide concentration, starting ventilation equipment or supplementing the air source; When the analysis result indicates that the pH value of the nutrient solution needs to be adjusted, the nutrient solution mixing device is started and the mixing ratio is adjusted.

7. A remote management device for fish and vegetable symbiosis based on digital twin, characterized in that: The device comprises: A model building unit is used to obtain physical parameter data of the physical fish pond and physical plant shed in the aquaponics system, and create twin models corresponding to the fish school, fish pond, hydroponic plants, plant shed and physical equipment of the aquaponics system based on the physical parameter data to obtain the aquaponics digital twin platform; A data acquisition unit, configured to acquire, in real time, fish farming parameters, plant growth parameters, and equipment operating parameters collected by sensors and cameras deployed in the aquaponics system, and upload the data to a cloud server; a data analysis unit, configured to analyze the fish farming parameters, plant growth parameters, and equipment operation parameters through the aquaponics digital twin platform to obtain analysis results; Wherein, the data analysis unit is further used for: When the fish pond environmental parameters meet the preset growth environment range, the fish health status of the fish in different fish ponds is analyzed and the growth trend of the fish is predicted based on the fish image data combined with the feeding amount, wherein the fish pond environmental parameters include fish pond water temperature, dissolved oxygen concentration, water quality parameters, and pH value; When the plant environmental parameters meet the preset growth environment range, the plant growth status of hydroponic plants in different planting racks is analyzed and the plant growth trend and yield are predicted based on the plant image data combined with the biological characteristics of the plants. The plant environmental parameters include plant shed temperature, humidity, light intensity, carbon dioxide concentration, and nutrient solution pH value. When there are abnormalities in the health status of fish or plant growth, the historical disease database is called for matching and a treatment plan is obtained; Analyzing the operating parameters of the device, detecting whether the physical device has an abnormality, obtaining an abnormality treatment plan, and using the treatment plan and / or the abnormality treatment plan as the analysis result; Among them, the fish image data combined with the feeding amount are used to analyze the health status of fish in different fish ponds and predict the growth trend of fish schools, including: Preprocessing the fish image data and extracting image features to obtain the fish appearance characteristics; The fish appearance characteristics are input into the constructed health status analysis model and fish growth prediction model respectively to obtain health status analysis results and fish growth prediction results respectively; The health status analysis model is constructed based on a dual-branch multi-task learning model, which inputs the fish's appearance characteristics into the health status analysis model and outputs health status analysis results including health level and disease type; The fish growth prediction model is constructed based on the LSTM model, which inputs the fish's appearance characteristics, fish environmental parameters and historical time series into the fish growth prediction model and outputs the fish growth prediction results of weight growth rate and body length change in the next 7 days; Among them, the plant growth status of hydroponic plants in different planting racks is analyzed based on plant image data combined with the biological characteristics of the plants, and the plant growth trend and yield are predicted, including: Perform image segmentation and image feature extraction on plant image data to obtain multispectral features; Inputting the multispectral features into the constructed growth state analysis model and plant growth prediction model respectively to obtain growth state analysis results and plant growth prediction results respectively; The growth status analysis model is constructed based on a random forest classifier and an isolation forest algorithm, multispectral features and plant environmental parameters are input into the growth status analysis model, and the growth status analysis results including growth status grade and plant disease type are output; The plant growth prediction model is built based on the LSTM-Transformer hybrid model, which combines multispectral features with environmental parameter trends to input the plant growth prediction model and outputs plant growth prediction results for the next 30 days, including LAI and yield. A remote management unit is configured to trigger an early warning mechanism and generate control instructions based on the analysis results, and dynamically adjust the operating status of corresponding physical devices in the aquaponics system according to the control instructions, thereby achieving remote collaborative management of fish farming, plant growth, and equipment operation; The device also includes a fish recognition model constructed based on a dual-branch multi-task learning model to achieve the dual-task requirements of growth stage classification and disease state identification. The fish recognition model uses ConvNeXt-L as a shared feature extractor and is combined with a dynamic sparse attention mechanism; wherein, the fish recognition model includes a growth stage classification branch and a disease state identification branch, the growth stage classification branch includes a Non-local Blocks module for capturing the global morphological characteristics of the fish body, the disease state identification branch includes a YOLOv8-P2 structure for outputting the disease type and lesion location, the fish recognition model separates growth-related features from disease features through a gradient inversion layer, embeds a BiFPN structure in the backbone network, and uses depth-separable convolution for model lightweight processing.

8. A remote management device for aquaponics based on digital twins, characterized in that: The system comprises a processor, a memory, and a computer program stored in the memory, wherein the computer program is executed by the processor to implement the steps of a remote management method of aquaponics based on digital twins as described in any one of claims 1 to 6.

9. A computer program product, characterized in that The method comprises a computer program / instruction, which, when executed by a processor, implements the steps of a remote management method of aquaponics based on digital twins as described in any one of claims 1 to 6.

Images