A fish disease detection method and device, computer equipment and storage medium

Abstract

The application discloses a fish disease detection method and device, computer equipment and a storage medium, belonging to the field of artificial intelligence. First, by acquiring and analyzing the image of the farm, target recognition is performed to determine the distribution position of the fish body, and then the image is segmented to obtain a plurality of individual fish body images. The fish body image is input into the fish disease detection model, the fish body carrying the disease is identified, and then the fish disease image is determined. Then, the image of the farm is acquired again, and image matching is performed to determine the specific position of the diseased fish. The system continuously monitors the position to obtain the motion path of the fish disease, and outputs the final fish disease image and its motion path to the client. The application also relates to the field of blockchain technology, and the fish body image is stored on the blockchain network. The application utilizes image recognition, target detection and deep learning model to realize automatic fish disease monitoring and tracking, and improves the efficiency and accuracy of fish disease prevention and control.

Description

Technical Field

[0001] The present application belongs to the field of artificial intelligence technology, and specifically relates to a fish disease detection method, device, computer equipment and storage medium. Background Art

[0002] Aquaculture is one of the fastest-growing industries globally. Artificial intelligence technology has been widely researched and applied in a variety of sub-sectors, including water quality monitoring, intelligent feeding, and remote temperature control. However, breakthroughs in fish disease detection, both in terms of quality and speed, have yet to be achieved. Summer is the peak season for fish reproduction in aquaculture, but due to high temperatures and heavy rainfall, fish pathogens also multiply, leading to various farmed fish diseases. Viral fish diseases, in particular, are highly contagious and, if unintentionally spread, can cause mass fish mortality, significantly impacting the aquaculture industry.

[0003] Farmed fish disease detection is extremely important for the underwriting and claims settlement of aquaculture insurance. However, the current popular method of fish disease detection in the industry mainly relies on manual processing methods such as visual observation of fry activity and feeding status, body color and body shape. However, the number of fry in fish ponds is huge, and manual visual observation is time-consuming and labor-intensive, and it is difficult to ensure that fish diseases are discovered in the first place. This often leads to inaccurate risk assessments when aquaculture insurance is underwritten, and disputes arise due to insufficient evidence or delays in claims settlement, affecting the efficiency of insurance services and the protection of farmers' rights and interests. Summary of the Invention

[0004] The purpose of the embodiments of the present application is to propose a fish disease detection method, device, computer equipment and storage medium to solve the technical problems that the existing farmed fish disease detection adopts a manual processing method of observing the condition of fry with the naked eye, which is inefficient, time-consuming and labor-intensive, and difficult to ensure that fish diseases are discovered in the first place, resulting in inaccurate risk assessment when underwriting and claiming aquaculture insurance, as well as disputes caused by insufficient evidence or delays, affecting the efficiency of insurance services and the protection of the rights and interests of farmers.

[0005] In order to solve the above technical problems, the present invention provides a method for detecting fish diseases, which adopts the following technical solutions:

[0006] A fish disease detection method comprising:

[0007] Obtaining a detection instruction, obtaining an image of the breeding site according to the detection instruction to obtain a first breeding site image, and performing target recognition on the first breeding site image to obtain a first target recognition result;

[0008] Performing image segmentation on the first breeding site image according to the first target recognition result to obtain a plurality of fish body images;

[0009] Import several fish images into the pre-trained fish disease detection model to identify fish carrying fish diseases;

[0010] Among a plurality of fish body images, identifying fish body images carrying fish disease bodies, and obtaining fish disease body images;

[0011] Acquire the breeding site image again to obtain a second breeding site image, and perform image matching on the second breeding site image and the fish disease body image to determine the position of the fish disease body in the second breeding site image;

[0012] The position of the diseased fish in the image of the second breeding site is monitored to obtain a movement path of the diseased fish, and the image of the diseased fish and the movement path of the diseased fish are output to the client.

[0013] Furthermore, a plurality of cameras are arranged around the breeding site, each camera is responsible for photographing a portion of the breeding site, obtaining a detection instruction, and obtaining an image of the breeding site according to the detection instruction. The steps of obtaining the first breeding site image specifically include:

[0014] Receive detection instructions and instruct several cameras to shoot the breeding site according to the detection instructions;

[0015] Acquire images of the breeding ground taken by several cameras to obtain several regional images of the breeding ground;

[0016] Perform image fusion on several breeding site area images to obtain a first breeding site image.

[0017] Furthermore, each camera is connected to an edge server, and each edge server is connected to a cloud server. After obtaining the images of the breeding site taken by the plurality of cameras and obtaining the images of the breeding site area, the method further includes:

[0018] For each breeding site area image, image feature extraction is performed on the breeding site area image to obtain regional image features;

[0019] Performing image block processing on the breeding site area image according to the regional image features to obtain a number of image sub-blocks;

[0020] Transmitting several image sub-blocks from the edge server to the cloud server in sequence;

[0021] On the cloud server, image reconstruction is performed based on the image features of the image sub-blocks to obtain an image of the breeding site area.

[0022] Furthermore, before the step of importing the plurality of fish images into the pre-trained fish disease detection model to identify fish with fish diseases, the method further includes:

[0023] acquire historical fish body images, wherein the historical fish body images include historical fish disease body images and historical normal fish body images;

[0024] construct a data sample set by taking the historical fish disease body images as positive sample data and taking the historical normal fish body images as negative sample data;

[0025] import the data sample set into a preset image classification model, so that the image classification model learns image features in the data sample set, and obtain a pre-trained fish disease detection model.

[0026] Further, the data sample set includes a training data set and a validation data set, and the step of importing the data sample set into the preset image classification model so that the image classification model learns image features in the data sample set to obtain the pre-trained fish disease detection model specifically includes:

[0027] extract image features of sample images in the training data set to obtain training sample image features;

[0028] perform associated learning on the training sample image features and a preset classification label in the image classification model to obtain an initial fish body classification model;

[0029] perform model validation on the initial fish body classification model using the validation data set, and iteratively train the initial fish body classification model based on the model validation result until the model is fitted to obtain the fish disease detection model.

[0030] Further, the step of again acquiring the aquaculture site image to obtain a second aquaculture site image and performing image matching on the second aquaculture site image and the fish disease body image to determine the position of the fish disease body in the second aquaculture site image specifically includes:

[0031] again acquire the aquaculture site image to obtain a second aquaculture site image, and perform target recognition on the second aquaculture site image to obtain a second target recognition result;

[0032] perform image segmentation on the second aquaculture site image according to the second target recognition result to obtain a plurality of fish body images to be detected;

[0033] perform image feature extraction on the fish disease body image to obtain first fish body features, and perform image feature extraction on the plurality of fish body images to be detected to obtain second fish body features;

[0034] calculate the similarity of the first fish body features and the second fish body features;

[0035] when the similarity of the first fish body features and the second fish body features is greater than or equal to a preset similarity threshold, determine a fish body matching the fish disease body in the second aquaculture site image to obtain a target fish body;

[0036] A position coordinate system of the breeding site is constructed, and the position of the target fish body is determined in the position coordinate system to obtain the position of the diseased fish body in the second breeding site image.

[0037] Furthermore, the steps of monitoring the position of the diseased fish in the image of the second breeding site, obtaining the movement path of the diseased fish, and outputting the image of the diseased fish and the movement path of the diseased fish to the client specifically include:

[0038] Determine the position of the diseased fish in the second breeding site image as a first position;

[0039] Acquire images of the breeding site according to a preset time period to obtain a plurality of third breeding site images;

[0040] Determine the position of the target fish in the position coordinate system, and obtain the position of the diseased fish in the images of the plurality of third breeding sites;

[0041] determining the position of the diseased fish in the third breeding site image as a second position;

[0042] Connect the first position and the second position to obtain the movement path of the fish;

[0043] The diseased fish image and the diseased fish movement path are output to the client.

[0044] In order to solve the above technical problems, the present application also provides a fish disease detection device, which adopts the following technical solution:

[0045] A fish disease detection device, comprising:

[0046] a first image acquisition module, configured to acquire a detection instruction, acquire an image of the breeding site according to the detection instruction, obtain a first breeding site image, and perform target recognition on the first breeding site image to obtain a first target recognition result;

[0047] An image segmentation module, configured to segment the first aquaculture site image according to the first target recognition result to obtain a plurality of fish body images;

[0048] The fish disease detection module is used to import several fish images into the pre-trained fish disease detection model to identify fish carrying fish diseases;

[0049] The fish diseased body image matching module is used to identify the fish body image carrying the fish diseased body among a plurality of fish body images and obtain the fish diseased body image;

[0050] a second image acquisition module, configured to acquire the aquaculture site image again to obtain a second aquaculture site image, and perform image matching between the second aquaculture site image and the fish diseased body image to determine the position of the fish diseased body in the second aquaculture site image;

[0051] The diseased fish monitoring module is used to monitor the position of the diseased fish in the image of the second breeding site, obtain the movement path of the diseased fish, and output the diseased fish image and the movement path of the diseased fish to the client.

[0052] In order to solve the above technical problems, the embodiment of the present application further provides a computer device, which adopts the following technical solution:

[0053] A computer device comprises a memory and a processor, wherein the memory stores computer-readable instructions, and the processor implements the steps of any one of the above-mentioned fish disease detection methods when executing the computer-readable instructions.

[0054] In order to solve the above technical problems, the embodiment of the present application further provides a computer-readable storage medium, which adopts the following technical solution:

[0055] A computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of any one of the above-mentioned fish disease detection methods.

[0056] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0057] This application discloses a fish disease detection method, device, computer equipment, and storage medium, belonging to the field of artificial intelligence technology. Through image recognition, target detection, and deep learning models, this application automatically identifies and tracks the symptoms of farmed fish. This effectively solves the inefficiency, time-consuming, and labor-intensive problem of traditional manual observation of fry, achieves real-time and accurate monitoring of fish diseases, and ensures early detection of fish diseases. Furthermore, through automatic monitoring of the location of diseased fish and tracking of their motion paths, the system can provide accurate evidence of fish diseases, improving the accuracy of risk assessments in aquaculture insurance, avoiding claims disputes caused by delayed discovery or insufficient evidence, thereby improving insurance service efficiency and further protecting the rights and interests of farmers. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the solutions in this application, a brief introduction will be given below to the drawings required for use in the description of the embodiments of this application. Obviously, the drawings described below are some embodiments of this application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0059] Figure 1 shows an exemplary system architecture diagram in which the present application can be applied;

[0060] Figure 2 A flow chart showing an embodiment of a fish disease detection method according to the present application is shown;

[0061] Figure 3 A schematic structural diagram of an embodiment of a fish disease detection device according to the present application is shown;

[0062] Figure 4 A schematic structural diagram of an embodiment of a computer device according to the present application is shown. DETAILED DESCRIPTION

[0063] Unless otherwise defined, all technical and scientific terms used herein have the same meanings as commonly understood by those skilled in the art to which this application belongs. The terms used in the specification of the application are for the purpose of describing specific embodiments only and are not intended to limit this application. The terms "including" and "having" and any variations thereof in the specification and claims of this application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of this application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0064] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute an independent or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0065] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0066] like Figure 1 As shown, system architecture 100 may include a terminal device 101, a network 102, and a server 103. Terminal device 101 may be a laptop computer 1011, a tablet computer 1012, or a mobile phone 1013. Network 102 is a medium for providing a communication link between terminal device 101 and server 103. Network 102 may include various connection types, such as wired or wireless communication links or fiber optic cables.

[0067] The user can use the terminal device 101 to interact with the server 103 via the network 102 to receive or send messages, etc. Various communication client applications can be installed on the terminal device 101, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0068] The terminal device 101 can be various electronic devices with a display screen and supporting web browsing. In addition to the laptop computer 1011, tablet computer 1012 or mobile phone 1013, the terminal device 101 can also be an e-book reader, an MP3 player (Moving Picture Experts Group Audio Layer III), an MP4 (Moving Picture Experts Group Audio Layer IV) player, a laptop computer and a desktop computer, etc.

[0069] The server 103 may be a server that provides various services, such as a background server that provides support for web pages displayed on the terminal device 101 .

[0070] It should be noted that the fish disease detection method provided in the embodiment of the present application is generally executed by a server / terminal device, and accordingly, the fish disease detection device is generally set in the server / terminal device.

[0071] It should be understood that Figure 1 The numbers of terminal devices, networks and servers in the embodiment are merely illustrative. The above system may have any number of terminal devices, networks and servers according to implementation requirements.

[0072] Continue to refer Figure 2 , shows a flow chart of an embodiment of a method for detecting fish diseases according to the present application. The method for detecting fish diseases comprises the following steps:

[0073] S201, obtaining a detection instruction, obtaining a breeding site image according to the detection instruction, obtaining a first breeding site image, and performing target recognition on the first breeding site image to obtain a first target recognition result.

[0074] Specifically, the system first receives a detection instruction, which is usually issued by the farm's monitoring personnel or automated control system. Once the instruction is issued, the system will use cameras or other imaging equipment to collect real-time images of the farm for image analysis and processing. The acquired image is called the first farm image and contains the overall distribution of fish in the current pond or breeding area. The system then performs target recognition on the image, mainly identifying the location and general outline of all fish. This step is achieved through computer vision technology, which may include edge detection, contour extraction, or other feature recognition methods to accurately distinguish fish from background or other unrelated objects.

[0075] Object recognition can be achieved using the YOLO neural network model. YOLO (You Only Look Once) is a real-time object detection model based on a convolutional neural network (CNN). The YOLO model is popular for its efficient object detection speed and high accuracy, making it suitable for real-time applications with limited computing resources, such as video surveillance, autonomous driving, and robotic vision. The YOLO model divides images of aquaculture sites into multiple grids and determines the presence of fish within each grid, as well as the fish's type and location. Through training, the YOLO model can accurately identify different types of fish, such as carp, crucian carp, and grass carp. Furthermore, by analyzing the behavioral characteristics of fish in aquaculture site images, the YOLO model can detect fish pathology and abnormal behavior, such as abnormal swimming speed, clustering, or dispersion.

[0076] S202: performing image segmentation on the first breeding site image according to the first target recognition result to obtain a plurality of fish body images.

[0077] Specifically, the system uses the results of target recognition to further refine the first aquaculture site image and perform image segmentation. The purpose of image segmentation is to separate the fish area in the entire image from the background to obtain a separate fish image. Segmentation can use traditional image processing algorithms such as threshold segmentation and region growing, or more advanced deep learning segmentation models such as U-Net and Mask R-CNN. These technologies can accurately identify the edges and contours of the fish, allowing each fish to be independently extracted as a separate image. These independent fish images will be used for disease detection in subsequent steps. Therefore, the accuracy of image segmentation directly affects the accuracy of fish disease detection.

[0078] S203: importing a plurality of fish images into a pre-trained fish disease detection model to identify fish carrying fish diseases.

[0079] Specifically, the system inputs multiple individual fish images extracted from image segmentation into a pre-trained fish disease detection model. This detection model, based on deep learning technology and trained with a large amount of labeled data, can identify the visual characteristics of specific fish diseases. For example, the model can detect disease characteristics such as abnormal scale color, skin lesions, and parasite attachment. The model can be constructed using structures such as convolutional neural networks (CNNs), which automatically identify diseased fish by extracting feature vectors and comparing them with known diseases. In this process, the model not only determines whether the fish is diseased but also outputs a probability score for the specific disease, providing a basis for identifying diseased fish images.

[0080] S204: Identify fish body images carrying fish disease among the plurality of fish body images to obtain fish disease body images.

[0081] Specifically, after the fish disease detection model identifies fish that may be carrying a disease, the system extracts these diseased fish images and generates diseased fish images. This step filters all test results, ensuring that the system only retains images of confirmed diseased fish and excludes images of healthy fish. Each diseased fish image includes detailed characteristics of the fish, such as disease type, location, and symptoms, to facilitate further analysis. Diseased fish images are not only easy to save as records, but also can be further analyzed to assess the severity and changing trends of symptoms in diseased fish, providing reliable data support for tracking and intervention.

[0082] S205 , acquiring the breeding site image again to obtain a second breeding site image, and performing image matching between the second breeding site image and the fish disease body image to determine the position of the fish disease body in the second breeding site image.

[0083] Specifically, after identifying the image of the diseased fish, the system will re-acquire a real-time ground image of the farm after a period of time—this is known as the second farm image—to update the actual location of the diseased fish. The system will then match the newly acquired second farm image with the previous diseased fish image, identifying the real-time location of the diseased fish by comparing positional features. This matching process relies on algorithms such as image feature point matching and pattern recognition to ensure accurate tracking of the diseased fish's position in the new image. This image matching method enables the system to track diseased fish in real time in dynamic aquatic environments, avoiding positioning errors caused by the fish's movement and improving monitoring accuracy.

[0084] S206: Monitor the position of the diseased fish in the image of the second breeding site to obtain a movement path of the diseased fish, and output the image of the diseased fish and the movement path of the diseased fish to the client.

[0085] Specifically, once the specific location of the sick fish is determined, the system will initiate real-time tracking and monitoring of the sick fish to obtain its movement path. By continuously capturing the sick fish's location, the system can depict the fish's swimming trajectory and record its movement patterns. This data can provide information on the sick fish's behavior, such as whether it has become isolated from the group or whether it exhibits abnormal swimming movements, supporting disease diagnosis. The system also sends information such as images of the sick fish and their movement paths to the client, allowing managers to view the real-time dynamics of the sick fish on the client device. This data output method enables aquaculture managers to take timely measures, such as isolating the sick fish or administering medication, improving the efficiency of disease treatment and reducing the risk of disease spread.

[0086] During the aquaculture insurance application and claims process, this fish disease detection solution can provide timely and accurate diseased fish identification and data support. For example, before issuing insurance, insurers need to assess the risk profile of aquaculture farms. Traditionally, obtaining accurate information about diseased fish has been difficult, leading to highly subjective assessments. With this solution, the system automatically captures images of fish at the farm and uses a pre-trained model to detect diseased fish in real time, transmitting the images and data to the insurance company. This allows insurers to conduct accurate risk assessments based on real-time detection data, determine reasonable insurance premiums based on the type, number, and distribution of fish diseases, and ensure fair and targeted insurance plans.

[0087] During the claims process, this solution can also help insurance companies and fish farmers verify the occurrence and progression of fish diseases, avoiding disputes caused by insufficient evidence. When a disease outbreak occurs in a fish population, the system monitors the distribution and movement paths of diseased fish in real time, capturing images and movement data of the diseased fish as evidence. Insurance company claims adjusters can use this data to determine whether the diseased fish caused the death or loss of the insured risk, clarifying the cause of the disease and attribution of liability, thus avoiding disputes. The system also records the evolution and spread of disease symptoms in diseased fish, providing a basis for determining whether the spread of the disease is controllable, thereby improving the efficiency and transparency of claims processing.

[0088] For fish farmers, the system also effectively protects their rights. By monitoring sick fish in real time, farmers can receive timely disease warnings and take appropriate measures to mitigate losses. In the event of a large-scale disease outbreak, farmers can submit diseased fish detection data directly to insurance companies as strong evidence for claims, ensuring a faster and smoother claims process. This digitized, automated disease detection method significantly shortens claim review times and avoids claims disputes caused by a lack of diseased fish monitoring records, enabling farmers to receive more timely compensation.

[0089] In the above embodiments, the present application realizes automatic identification and tracking of the disease conditions of farmed fish through image recognition, target detection and deep learning models, effectively solving the inefficient, time-consuming and labor-intensive problems of traditional manual observation of fish fry, and realizing real-time and accurate monitoring of fish diseases, ensuring the early detection of fish diseases. Moreover, through automatic monitoring of the position of diseased fish and tracking of movement paths, the system can provide accurate evidence of fish diseases, thereby improving the accuracy of risk assessment in aquaculture insurance, avoiding claims disputes caused by delayed discovery or insufficient evidence, thereby improving the efficiency of insurance services and further protecting the rights and interests of farmers.

[0090] Furthermore, a plurality of cameras are arranged around the breeding site, each camera is responsible for photographing a portion of the breeding site, obtaining a detection instruction, and obtaining an image of the breeding site according to the detection instruction. The steps of obtaining the first breeding site image specifically include:

[0091] Receive detection instructions and instruct several cameras to shoot the breeding site according to the detection instructions;

[0092] Acquire images of the breeding ground taken by several cameras to obtain several regional images of the breeding ground;

[0093] Perform image fusion on several breeding site area images to obtain a first breeding site image.

[0094] In this embodiment, the system first receives a detection instruction, and controls several cameras around the breeding site according to the instruction to synchronously shoot different areas. Each camera covers a part of the breeding site, ensuring that the areas shot by each camera can be seamlessly connected to achieve full-area monitoring of the breeding site. Subsequently, the system collects the images shot by each camera, that is, the regional images of the breeding site, which correspond to the various areas covered by the camera. Then, the system performs image fusion processing on these regional images, and generates a complete first breeding site image through steps such as splicing, alignment, and correction. Image fusion technology ensures that the images of each region are consistent in brightness, contrast, and geometric alignment to form a unified panoramic picture, which is convenient for target detection and fish disease identification processing. This panoramic image generation method is suitable for large-scale farms and can clearly show the distribution of fish.

[0095] Through the above steps, the system can generate high-quality images covering the entire farm, making the monitoring perspective during diseased fish detection more comprehensive and helping to improve the accuracy and efficiency of detection.

[0096] Furthermore, each camera is connected to an edge server, and each edge server is connected to a cloud server. After obtaining the images of the breeding site taken by the plurality of cameras and obtaining the images of the breeding site area, the method further includes:

[0097] For each breeding site area image, image feature extraction is performed on the breeding site area image to obtain regional image features;

[0098] Performing image block processing on the breeding site area image according to the regional image features to obtain a number of image sub-blocks;

[0099] Transmitting several image sub-blocks from the edge server to the cloud server in sequence;

[0100] On the cloud server, image reconstruction is performed based on the image features of the image sub-blocks to obtain an image of the breeding site area.

[0101] In this embodiment, each camera transmits the captured image of the breeding site area to the edge server connected to it. The edge server pre-processes the received area image, first extracting the image's feature information, which can be key points, edge information, etc. Then, the system blocks the breeding site area image according to the extracted image features, decomposing it into several image sub-blocks. The block operation can reduce the amount of data transmitted each time, alleviate network transmission pressure, and enable large-scale images to be transmitted to the cloud server more efficiently. Next, each edge server transmits the processed image sub-blocks to the cloud server one by one. After receiving the image, the cloud server reassembles the image based on the image features of each image sub-block, reconstructs the image, and restores the complete breeding site area image. During image reconstruction, the cloud server uses the image feature information to ensure the correct splicing and fusion between the image sub-blocks, ensuring that the reconstructed area image is of high quality and coherence.

[0102] Through the above steps, the system can efficiently complete image transmission and reconstruction, significantly reducing bandwidth requirements while ensuring image quality, and improving the real-time performance and reliability of data processing.

[0103] Furthermore, before the step of importing the plurality of fish images into the pre-trained fish disease detection model to identify fish with fish diseases, the method further includes:

[0104] Acquire historical fish body images, wherein the historical fish body images include historical diseased fish body images and historical normal fish body images;

[0105] The historical diseased fish images are used as positive sample data, and the historical normal fish images are used as negative sample data to construct a data sample set;

[0106] The data sample set is imported into a preset image classification model, so that the image classification model learns the image features in the data sample set to obtain a pre-trained fish disease detection model.

[0107] In this embodiment, in order to improve the accuracy of the fish disease detection model, the system first trains the model with historical data before the formal detection. First, the system obtains historical fish images from the database. These images include historical diseased fish images and normal fish images that have been labeled. The system labels the diseased fish images as positive samples and the normal fish images as negative samples to construct a set of labeled data sample sets. Then, this data sample set is input into the preset image classification model for training. By learning the feature differences between these positive and negative samples, the image classification model can gradually grasp the typical features of diseased fish, such as color, morphology, and lesion location. During the training process, the model optimizes the parameters through multiple iterations, thereby having the ability to identify lesions in new images. The model pre-trained in this way becomes a pre-trained fish disease detection model, which can recognize and judge the lesion features in future fish images.

[0108] An image classification model is an intelligent model built based on a deep learning algorithm. It can learn image features and classify images into different categories using a large number of labeled samples. In fish disease detection applications, image classification models are trained to distinguish between "sick fish" and "healthy fish." By comparing positive samples (sick fish images) and negative samples (normal fish images), the model learns the typical features of the two types of images, such as color, shape, texture, and other unique features of diseased areas. During training, the model continuously adjusts its internal parameters to optimize the recognition of fish disease features. Ultimately, the model can automatically extract and analyze features from the input image to determine whether the fish is diseased, thereby achieving fast and accurate classification during the fish disease detection process.

[0109] Through the above steps, the system can obtain a fish disease detection model with good detection accuracy, thereby effectively improving the accuracy of diseased fish identification and ensuring the reliability of the test results.

[0110] Furthermore, the data sample set includes a training data set and a validation data set. The steps of importing the data sample set into a preset image classification model so that the image classification model learns the image features in the data sample set to obtain a pre-trained fish disease detection model specifically include:

[0111] Extracting image features of sample images in the training data set to obtain training sample image features;

[0112] In the image classification model, the training sample image features and the preset classification labels are associated with each other to obtain the initial fish classification model;

[0113] The initial fish classification model was verified using the validation dataset, and the initial fish classification model was iteratively trained based on the model verification results until the model was fitted to obtain a fish disease detection model.

[0114] In this embodiment, the system first extracts the features of each sample image from the training data set, such as texture, shape, color, etc., to form a training sample image feature set, and assigns corresponding classification labels (such as "sick fish" or "healthy fish") to these samples. Then, the training sample image feature set and classification labels are imported into the pre-set image classification model for associated learning, generating a preliminary fish body classification model, i.e., an initial fish body classification model. This model will then be validated by a validation data set, which did not participate in the initial training, so as to more truly reflect the classification ability of the model. The system evaluates the accuracy of the model according to the validation result, and further iteratively trains the model by adjusting the model parameters to improve its recognition ability of fish disease characteristics. This process continues until the model achieves the desired fitting effect on the validation data set, at which time the final fish disease detection model is generated. This model has the ability to efficiently and accurately identify sick fish from new images.

[0115] During the associated learning, the system will correspond the training sample image feature set (such as the feature vector containing the color, shape, texture, etc. of the fish body) with the corresponding classification label (such as "sick fish" or "healthy fish") one by one, and import it into the image classification model. During the learning process, the model calculates the relationship between the input image features and the classification labels, and gradually identifies the feature differences of different categories of images. For example, in the early stage of training, the model randomly initializes its parameters, and each time the image features are input into the model, the model generates a predicted label, which is then compared with the actual label (classification label) to calculate the error. The error adjusts the model parameters through backpropagation, so that the model can gradually reduce the misclassification. After multiple rounds of training, the model learns the patterns between fish disease characteristics and labels, so that it can accurately predict the health status of the fish body according to the image features. This process of associated learning enables the model to ultimately identify sick fish or healthy fish based on fish body characteristics.

[0116] Through the above steps, the system obtains a fish disease detection model that has been comprehensively trained and validated, enabling the model to achieve high recognition accuracy in actual detection and ensuring the reliability and stability of the sick fish identification results.

[0117] Further, the second aquaculture site image is obtained again, and the second aquaculture site image and the fish disease image are matched to determine the position of the fish disease in the second aquaculture site image, which specifically includes:

[0118] The second aquaculture site image is obtained again, and target recognition is performed on the second aquaculture site image to obtain a second target recognition result;

[0119] According to the second target recognition result, the second aquaculture site image is segmented to obtain a plurality of fish body images to be detected;

[0120] Performing image feature extraction on the diseased fish image to obtain a first fish body feature, and performing image feature extraction on a plurality of fish body images to be detected to obtain a second fish body feature;

[0121] Calculating the similarity between the first fish body feature and the second fish body feature;

[0122] When the similarity between the first fish body feature and the second fish body feature is greater than or equal to a preset similarity threshold, a fish body matching the diseased fish body in the second breeding site image is determined to obtain a target fish body;

[0123] A position coordinate system of the breeding site is constructed, and the position of the target fish body is determined in the position coordinate system to obtain the position of the diseased fish body in the second breeding site image.

[0124] In this embodiment, the system first uses a camera to capture an overall image of the farm again, namely the second farm image, and performs target recognition on this image to determine the specific distribution and location of the fish within the farm, forming a second target recognition result. Based on this result, the system performs image segmentation on the second farm image and extracts several images of fish to be detected. Next, the system performs feature extraction on the images of the initially detected diseased fish. The extracted features include unique texture, color, or shape features, forming a first set of fish features. Simultaneously, features are extracted from each of the images of the fish to be detected, forming a second set of fish features. The system then calculates the similarity between the first fish feature and each second fish feature to identify the most matching fish location. If the similarity reaches or exceeds a set threshold, the corresponding diseased fish is considered to have been found and marked as the target fish. The system then determines the specific location of the target fish by constructing a coordinate system for the farm, thereby obtaining the precise location of the diseased fish in the second farm image.

[0125] Through the above steps, the system can efficiently locate and track the location of diseased fish in the farm, making subsequent monitoring and prevention and control measures more targeted, and improving the accuracy and efficiency of fish disease detection and management.

[0126] Furthermore, the steps of monitoring the position of the diseased fish in the image of the second breeding site, obtaining the movement path of the diseased fish, and outputting the image of the diseased fish and the movement path of the diseased fish to the client specifically include:

[0127] Determine the position of the diseased fish in the second breeding site image as a first position;

[0128] Acquire images of the breeding site according to a preset time period to obtain a plurality of third breeding site images;

[0129] Determine the position of the target fish in the position coordinate system, and obtain the position of the diseased fish in the images of the plurality of third breeding sites;

[0130] determining the position of the diseased fish in the third breeding site image as a second position;

[0131] Connect the first position and the second position to obtain the movement path of the fish;

[0132] The diseased fish image and the diseased fish movement path are output to the client.

[0133] In this embodiment, the system first marks the initial position of the identified diseased fish in the second aquaculture site image as the first position. Next, the system continuously acquires new aquaculture site images at a preset time period (e.g., every few seconds), referred to as third aquaculture site images, and tracks the position of each third aquaculture site image. Through coordinate matching, the system marks the position of the diseased fish in these images at each time point in a position coordinate system and records its relative coordinates. This provides the position information of the diseased fish at each time point, i.e., the second position. The system then sequentially connects the positions of the diseased fish at all time points to generate a movement path of the diseased fish, thereby reflecting the dynamic movement trajectory of the diseased fish in the aquaculture site. Using this path information, the system can accurately track the movement range of diseased fish, helping aquaculture managers understand the distribution and spread of diseased fish in the aquaculture environment. Finally, the system transmits the complete data of the diseased fish image and movement path to the client, allowing managers to view it in real time and take necessary measures.

[0134] Through the above steps, the system can achieve accurate tracking and path monitoring of diseased fish, making it easy to timely grasp the activity status of diseased fish, helping to take targeted intervention measures and reduce the risk of fish disease spread.

[0135] In the above embodiment, the present application discloses a fish disease detection method, which belongs to the field of artificial intelligence technology. This application uses image recognition, target detection, and deep learning models to automatically identify and track the symptoms of farmed fish. This effectively solves the inefficiency, time-consuming, and labor-intensive problem of traditional manual observation of fry, realizes real-time and accurate monitoring of fish diseases, ensures the early detection of fish diseases, and through automatic monitoring of the location of diseased fish and tracking of their movement paths, the system can provide accurate evidence of fish diseases, improve the accuracy of risk assessment in aquaculture insurance, avoid claims disputes caused by delayed discovery or insufficient evidence, thereby improving the efficiency of insurance services and further protecting the rights and interests of farmers.

[0136] In this embodiment, the fish disease detection method is run on an electronic device (e.g. Figure 1The server shown in the figure) can receive instructions or obtain data through a wired connection or a wireless connection. It should be noted that the above-mentioned wireless connection method may include but is not limited to 3G / 4G connection, WiFi connection, Bluetooth connection, WiMAX connection, Zigbee connection, UWB (ultra wideband) connection, and other wireless connection methods currently known or to be developed in the future.

[0137] It should be emphasized that in order to further ensure the privacy and security of the above-mentioned fish image information, the above-mentioned fish image information can also be stored in a node of a blockchain.

[0138] The blockchain referred to in this application is a new application model for computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanisms, and encryption algorithms. Blockchain is essentially a decentralized database, a series of data blocks generated using cryptographic methods. Each data block contains information about a batch of network transactions, which is used to verify the validity of this information (to prevent counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, the platform product service layer, and the application service layer.

[0139] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Artificial Intelligence (AI) is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to achieve optimal results.

[0140] Fundamental AI technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interaction systems, and mechatronics. AI software technologies primarily encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0141] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing related hardware via computer-readable instructions. The computer-readable instructions can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes in the above-described method embodiments. The aforementioned storage medium can be a non-volatile storage medium such as a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0142] It should be understood that although the steps in the flowcharts of the accompanying drawings are shown in sequence as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some of the steps in the flowcharts of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

[0143] Further references Figure 3 , as a response to the above Figure 2 The present application provides an embodiment of a fish disease detection device, which is similar to the embodiment of the present invention. Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0144] like Figure 3 As shown, the fish disease detection device 300 described in this embodiment includes:

[0145] The first image acquisition module 301 is used to obtain a detection instruction, obtain an image of the breeding site according to the detection instruction, obtain a first breeding site image, and perform target recognition on the first breeding site image to obtain a first target recognition result;

[0146] An image segmentation module 302 is configured to segment the first aquaculture site image according to the first target recognition result to obtain a plurality of fish body images;

[0147] The fish disease detection module 303 is used to import a plurality of fish body images into a pre-trained fish disease detection model to identify fish bodies carrying fish diseases;

[0148] The fish diseased body image matching module 304 is used to identify the fish body image carrying the fish diseased body among the multiple fish body images and obtain the fish diseased body image;

[0149] The second image acquisition module 305 is used to acquire the aquaculture site image again to obtain a second aquaculture site image, and perform image matching between the second aquaculture site image and the fish diseased body image to determine the location of the fish diseased body in the second aquaculture site image;

[0150] The diseased fish monitoring module 306 is used to monitor the position of the diseased fish in the second breeding site image, obtain the movement path of the diseased fish, and output the diseased fish image and the movement path of the diseased fish to the client.

[0151] Further, a plurality of cameras are arranged around the breeding site, each camera is responsible for shooting a part of the breeding site, and the first image acquisition module 301 is specifically configured to:

[0152] receive a detection instruction, and instruct the plurality of cameras to shoot the breeding site according to the detection instruction; acquire the breeding site images shot by the plurality of cameras to obtain a plurality of breeding site region images; and perform image fusion on the plurality of breeding site region images to obtain a first breeding site image.

[0153] Further, each camera is connected to an edge server, and each edge server is connected to a cloud server, and the first image acquisition module 301 is further configured to:

[0154] for each breeding site region image, perform image feature extraction on the breeding site region image to obtain a region image feature; perform image block processing on the breeding site region image according to the region image feature to obtain a plurality of image subblocks; sequentially transmit the plurality of image subblocks from the edge server to the cloud server; and based on the image features of the image subblocks, perform image reconstruction on the breeding site region image in the cloud server to obtain the breeding site region image.

[0155] Further, the fish disease detection device 300 further comprises a model training module, and the model training module is specifically configured to:

[0156] acquire historical fish body images, wherein the historical fish body images include historical fish disease body images and historical normal fish body images; construct a data sample set by taking the historical fish disease body images as positive sample data and taking the historical normal fish body images as negative sample data; and import the data sample set into a preset image classification model, so that the image classification model learns image features in the data sample set to obtain a pre-trained fish disease detection model.

[0157] Further, the data sample set includes a training data set and a verification data set, and the model training module is further configured to:

[0158] extract image features of sample images in the training data set to obtain training sample image features; in the image classification model, perform associated learning on the training sample image features and a preset classification label to obtain an initial fish body classification model; use the verification data set to perform model verification on the initial fish body classification model, and based on a model verification result, iteratively train the initial fish body classification model until the model is fitted to obtain the fish disease detection model.

[0159] Further, the second image acquisition module 305 is specifically configured to:

[0160] The breeding site image is acquired again to obtain a second breeding site image, and target recognition is performed on the second breeding site image to obtain a second target recognition result; the second breeding site image is segmented according to the second target recognition result to obtain a plurality of fish body images to be detected; image feature extraction is performed on the diseased fish image to obtain a first fish body feature, and image feature extraction is performed on the plurality of fish body images to be detected to obtain a second fish body feature; the similarity between the first fish body feature and the second fish body feature is calculated; when the similarity between the first fish body feature and the second fish body feature is greater than or equal to a preset similarity threshold, a fish body that matches the diseased fish body in the second breeding site image is determined to obtain a target fish body; a position coordinate system of the breeding site is constructed, and the position of the target fish body is determined in the position coordinate system to obtain the position of the diseased fish body in the second breeding site image.

[0161] Furthermore, the fish disease monitoring module 306 is specifically used to:

[0162] The position of the diseased fish body in the second breeding site image is determined as a first position; the breeding site image is acquired according to a preset time period to obtain a plurality of third breeding site images; the position of the target fish body is determined in a position coordinate system to obtain the position of the diseased fish body in the plurality of third breeding site images; the position of the diseased fish body in the third breeding site image is determined as a second position; the first position and the second position are connected to obtain a movement path of the diseased fish body; and the image of the diseased fish body and the movement path of the diseased fish body are output to a client.

[0163] In the above-mentioned embodiment, the present application discloses a fish disease detection device, which belongs to the field of artificial intelligence technology. This application uses image recognition, target detection, and deep learning models to automatically identify and track the symptoms of farmed fish. This effectively solves the inefficiency, time-consuming, and labor-intensive problem of traditional manual observation of fry, and achieves real-time and accurate monitoring of fish diseases, ensuring early detection of fish diseases. Moreover, through automatic monitoring of the location of diseased fish and tracking of their movement paths, the system can provide accurate evidence of fish diseases, improving the accuracy of risk assessment in aquaculture insurance, avoiding claims disputes caused by delayed discovery or insufficient evidence, thereby improving insurance service efficiency and further protecting the rights and interests of farmers.

[0164] To solve the above technical problems, the present application also provides a computer device. Figure 4 , Figure 4 This is a basic structural block diagram of the computer device in this embodiment.

[0165] The computer device 4 includes a memory 41, a processor 42, and a network interface 43 that are interconnected through a system bus. It should be noted that the figure only shows a computer device 4 with a memory 41, a processor 42, and a network interface 43, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to a microprocessor, an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA), a digital signal processor (DSP), an embedded device, etc.

[0166] The computer device may be a desktop computer, a notebook computer, a PDA computer, a cloud server, etc. The computer device may interact with the user via a keyboard, a mouse, a remote control, a touchpad, or a voice control device.

[0167] The memory 41 includes at least one type of readable storage medium, including flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic storage, magnetic disk, optical disk, etc. In some embodiments, the memory 41 can be an internal storage unit of the computer device 4, such as a hard disk or memory of the computer device 4. In other embodiments, the memory 41 can also be an external storage device of the computer device 4, such as a plug-in hard disk equipped on the computer device 4, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), etc. Of course, the memory 41 can also include both the internal storage unit of the computer device 4 and its external storage device. In this embodiment, the memory 41 is generally used to store the operating system and various application software installed on the computer device 4, such as computer-readable instructions of the fish disease detection method. In addition, the memory 41 can also be used to temporarily store various types of data that have been output or are to be output.

[0168] In some embodiments, the processor 42 can be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip. The processor 42 is generally used to control the overall operation of the computer device 4. In this embodiment, the processor 42 is used to execute computer-readable instructions or process data stored in the memory 41, such as computer-readable instructions for executing the fish disease detection method.

[0169] The network interface 43 may include a wireless network interface or a wired network interface. The network interface 43 is generally used to establish a communication connection between the computer device 4 and other electronic devices.

[0170] In the above embodiment, the present application discloses a computer device belonging to the field of artificial intelligence technology. This application uses image recognition, target detection, and deep learning models to automatically identify and track the symptoms of farmed fish. This effectively solves the inefficiency, time-consuming, and labor-intensive problem of traditional manual observation of fry, achieves real-time and accurate monitoring of fish diseases, and ensures early detection of fish diseases. Moreover, through automatic monitoring of the location of diseased fish and tracking of their movement paths, the system can provide accurate evidence of fish diseases, improve the accuracy of risk assessment in aquaculture insurance, avoid claims disputes caused by delayed discovery or insufficient evidence, thereby improving insurance service efficiency and further protecting the rights and interests of farmers.

[0171] The present application also provides another embodiment, namely, providing a computer-readable storage medium, which stores computer-readable instructions, and the computer-readable instructions can be executed by at least one processor to enable the at least one processor to perform the steps of the fish disease detection method as described above.

[0172] In the above embodiments, the present application discloses a computer-readable storage medium, which belongs to the field of artificial intelligence technology. This application uses image recognition, target detection, and deep learning models to automatically identify and track the symptoms of farmed fish. This effectively solves the inefficiency, time-consuming, and labor-intensive problem of traditional manual observation of fry, achieves real-time and accurate monitoring of fish diseases, and ensures early detection of fish diseases. Furthermore, through automatic monitoring of the location of diseased fish and tracking of their motion paths, the system can provide accurate evidence of fish diseases, improving the accuracy of risk assessments in aquaculture insurance, avoiding claims disputes caused by delayed discovery or insufficient evidence, thereby improving insurance service efficiency and further protecting the rights and interests of farmers.

[0173] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus the necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for enabling a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0174] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected via a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media, including storage devices.

[0175] Obviously, the embodiments described above are only some of the embodiments of the present application, rather than all of the embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application has been described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions described in the aforementioned specific embodiments, or to make equivalent replacements for some of the technical features therein. Any equivalent structure made using the contents of the present application specification and the accompanying drawings, directly or indirectly used in other related technical fields, is also within the scope of patent protection of the present application.

Claims

1. A method for detecting fish diseases, characterized in that: include: Obtaining a detection instruction, obtaining a breeding site image according to the detection instruction to obtain a first breeding site image, and performing target recognition on the first breeding site image to obtain a first target recognition result; performing image segmentation on the first aquaculture site image according to the first target recognition result to obtain a plurality of fish body images; Importing a plurality of fish images into a pre-trained fish disease detection model to identify fish carrying fish diseases; Identifying a fish body image carrying the diseased fish body from among the plurality of fish body images to obtain a diseased fish body image; Acquire the breeding site image again to obtain a second breeding site image, and perform image matching on the second breeding site image and the image of the fish disease body to determine the position of the fish disease body in the second breeding site image; monitoring the position of the diseased fish in the second aquaculture site image to obtain a movement path of the diseased fish, and outputting the image of the diseased fish and the movement path of the diseased fish to a client; The step of acquiring the aquaculture site image again to obtain a second aquaculture site image, and performing image matching on the second aquaculture site image and the fish disease body image to determine the position of the fish disease body in the second aquaculture site image specifically includes: Acquire the breeding site image again to obtain a second breeding site image, and perform target recognition on the second breeding site image to obtain a second target recognition result; performing image segmentation on the second aquaculture site image according to the second target recognition result to obtain a plurality of fish body images to be detected; Performing image feature extraction on the diseased fish image to obtain a first fish body feature, and performing image feature extraction on a plurality of the to-be-detected fish body images to obtain a second fish body feature; Calculating the similarity between the first fish body feature and the second fish body feature; When the similarity between the first fish body feature and the second fish body feature is greater than or equal to a preset similarity threshold, determining a fish body that matches the diseased fish body in the second breeding site image to obtain a target fish body; Constructing a position coordinate system of the breeding site, and determining the position of the target fish in the position coordinate system, and obtaining the position of the diseased fish in the second breeding site image; The step of monitoring the position of the diseased fish in the second breeding site image, obtaining a movement path of the diseased fish, and outputting the diseased fish image and the movement path of the diseased fish to the client specifically includes: Determining the position of the diseased fish in the second breeding site image as a first position; Acquire images of the breeding site according to a preset time period to obtain a plurality of third breeding site images; Determining the position of the target fish in the position coordinate system, and obtaining the position of the diseased fish in the plurality of images of the third breeding site; Determine the position of the diseased fish in the third breeding site image as a second position; Connecting the first position and the second position to obtain the movement path of the diseased fish; The diseased fish image and the diseased fish movement path are output to a client.

2. The fish disease detection method according to claim 1, wherein A plurality of cameras are arranged around the breeding site, each of which is responsible for photographing a portion of the breeding site. The steps of obtaining a detection instruction, obtaining an image of the breeding site according to the detection instruction, and obtaining a first breeding site image specifically include: receiving a detection instruction, and instructing a plurality of cameras to photograph the breeding ground according to the detection instruction; Acquire images of the breeding ground taken by a plurality of cameras to obtain a plurality of breeding ground area images; Perform image fusion on a plurality of the breeding site area images to obtain the first breeding site image.

3. The fish disease detection method according to claim 2, wherein Each of the cameras is connected to an edge server, and each of the edge servers is connected to a cloud server. After the step of obtaining the images of the breeding site taken by the cameras and obtaining the images of the breeding site area, the method further includes: For each breeding site area image, image feature extraction is performed on the breeding site area image to obtain area image features; Performing image block processing on the breeding site area image according to the regional image features to obtain a plurality of image sub-blocks; transmitting the plurality of image sub-blocks sequentially from the edge server to the cloud server; On the cloud server, image reconstruction is performed based on the image features of the image sub-blocks to obtain the breeding site area image.

4. The fish disease detection method according to claim 1, wherein Before the step of importing the plurality of fish images into the pre-trained fish disease detection model to identify fish with fish diseases, the method further includes: Acquire historical fish body images, wherein the historical fish body images include historical diseased fish body images and historical normal fish body images; The historical diseased fish images are used as positive sample data, and the historical normal fish images are used as negative sample data to construct a data sample set; The data sample set is imported into a preset image classification model, so that the image classification model learns the image features in the data sample set to obtain the pre-trained fish disease detection model.

5. The fish disease detection method according to claim 4, wherein The data sample set includes a training data set and a validation data set. The step of importing the data sample set into a preset image classification model so that the image classification model learns the image features in the data sample set to obtain the pre-trained fish disease detection model specifically includes: Extracting image features of sample images in the training data set to obtain training sample image features; In the image classification model, the training sample image features and the preset classification labels are associated with each other to obtain an initial fish classification model; The initial fish classification model is verified using the verification data set, and the initial fish classification model is iteratively trained based on the model verification result until the model is fitted to obtain the fish disease detection model.

6. A fish disease detection device, characterized in that: The fish disease detection device implements the steps of the fish disease detection method according to any one of claims 1 to 5, and the fish disease detection device comprises: a first image acquisition module, configured to acquire a detection instruction, acquire an image of a breeding site according to the detection instruction to obtain a first breeding site image, and perform target recognition on the first breeding site image to obtain a first target recognition result; an image segmentation module, configured to perform image segmentation on the first aquaculture site image according to the first target recognition result to obtain a plurality of fish body images; a fish disease detection module, configured to import a plurality of fish images into a pre-trained fish disease detection model to identify fish carrying fish diseases; a fish diseased body image matching module, configured to identify a fish body image carrying the fish diseased body from among a plurality of fish body images, and obtain a fish diseased body image; a second image acquisition module, configured to acquire the aquaculture site image again to obtain a second aquaculture site image, and perform image matching between the second aquaculture site image and the diseased fish image to determine a position of the diseased fish in the second aquaculture site image; The diseased fish monitoring module is used to monitor the position of the diseased fish in the second breeding site image, obtain the movement path of the diseased fish, and output the diseased fish image and the movement path of the diseased fish to the client.

7. A computer device, characterized in that: The method comprises a memory and a processor, wherein the memory stores computer-readable instructions, and the processor implements the steps of the fish disease detection method according to any one of claims 1 to 5 when executing the computer-readable instructions.

8. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of the fish disease detection method according to any one of claims 1 to 5.

Images