Breeding management system and method based on controllable animal husbandry

Abstract

The invention provides a breeding management system and method based on controllable animal husbandry, and the method comprises the steps: adjusting the environmental parameters of breeding equipment to affect a breeding region, positioning the position of the breeding region affected by the environment through the environmental parameters, configuring monitoring equipment to obtain a physiological image, and inputting the physiological image into a variety recognition model to determine an animal variety. And inputting the physiological image into a physiological index multi-task analysis model to identify a specific anomaly type, establishing an environment regulation and control strategy database to store a mapping relationship between the specific anomaly type corresponding to each animal variety and a target environment parameter, and matching a corresponding environment regulation and control strategy according to the specific anomaly type to adjust an environment parameter value of the breeding equipment. According to the invention, the accuracy of animal health monitoring and the scientificity of environment regulation and control can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of animal husbandry breeding management, and more specifically, to a breeding management system and method based on controllable animal husbandry. Background Art

[0002] Throughout the development of animal husbandry, breeding management has always been a key component. Traditional breeding management relies primarily on manual observation and empirical judgment. Staff members regularly check the animals' physiological condition, record their growth data, and adjust the breeding environment based on experience and established breeding standards. While this approach can meet basic breeding needs to a certain extent, it has many limitations. First, manual observation is easily influenced by subjective factors, resulting in inaccurate judgments of the animals' physiological condition. Second, the lack of real-time, accurate monitoring methods makes it difficult to detect animal health issues promptly, often escalating to the point where they become serious, increasing treatment costs and animal mortality. Furthermore, traditional methods lack scientific basis and precision when adjusting breeding environment parameters, making it impossible to tailor adjustments to the animals' specific physiological conditions and health risk levels, impacting animal growth efficiency and breeding profitability.

[0003] With the continuous advancement of science and technology, new technologies are being introduced into livestock breeding and management. For example, cameras and other monitoring equipment are used to capture animal images, and image recognition technology is used to determine the animal's breed and physiological condition. However, the application of these technologies is still in its early stages and has many shortcomings. For one thing, existing image recognition technology lacks high accuracy in identifying animal breeds and is easily affected by factors such as animal posture and lighting conditions. Furthermore, analysis of animal physiological characteristics is insufficiently in-depth, making it difficult to comprehensively assess animal health. Furthermore, existing technologies lack a close correlation between the adjustment of breeding environment parameters and the animal's health status, making it impossible to achieve precise regulation based on the animal's health risk level.

[0004] In implementing the embodiments of the present invention, existing technologies have at least the following problems or defects: Existing technologies cannot accurately identify animal breeds and physiological characteristics, making it difficult to comprehensively assess animal health; and when adjusting breeding environment parameters, existing technologies lack a close correlation with animal health, making it impossible to achieve precise regulation. These problems result in limited application of existing technologies in livestock breeding management, and they cannot meet the modern livestock industry's demand for efficient, precise, and intelligent breeding management. Summary of the Invention

[0005] The present invention provides a breeding management system and method based on controllable animal husbandry.

[0006] In a first aspect of the present invention, a breeding management method based on controlled animal husbandry is provided, comprising: S1. Adjust the environmental parameters of the breeding equipment so that the environmental parameters affect the breeding area to be managed; S2. Locate the breeding area affected by the environment based on the environmental parameters of the current breeding equipment; S3. Configuring monitoring equipment for the breeding equipment to obtain physiological images based on the location of the breeding area affected by the environment; S4. Inputting the physiological image into a species recognition model, and determining the animal species in the physiological image using the species recognition model; S5. Inputting the physiological image into a physiological index multi-task analysis model to identify a specific abnormality type in the physiological image; S6. Establishing an environmental control strategy database, wherein the environmental control strategy database is used to store a mapping relationship between a specific abnormality type corresponding to each animal species and a target environmental parameter; S7. Matching a corresponding environmental control strategy according to the specific abnormality type; S8. Adjust the environmental parameter values ​​of the breeding equipment according to the environmental control strategy.

[0007] Furthermore, the breed identification model includes a breed identification database, and the breed identification database includes detailed characteristics of each animal breed; Determining the animal species in the physiological image by using the species recognition model in S4 includes: Acquiring a first detail feature set in the physiological image; Comparing the first detailed feature set with the detailed features of each animal breed in the breed identification database to obtain the detailed features that overlap between the two and the number of the overlapping detailed features; Compare the number of overlapping detail features with the overlapping detail feature threshold, and rank the number of overlapping detail features corresponding to various types of animal breeds. If the number of overlapping detail features between the physiological image and a certain animal breed exceeds the overlapping detail feature threshold and the number of overlapping detail features is the largest, then output that the animal breed in the physiological image is the animal breed.

[0008] Furthermore, the training method of the variety recognition model includes: Construct the first training dataset, The first training data set includes physiological images and corresponding animal species labels; Inputting the first training data set into the breed recognition model to perform animal breed analysis, and determining a first loss value based on a difference between the animal breed label and the breed analysis result; The breed recognition model is iteratively updated and trained according to the first loss value, and the trained breed recognition model is used to determine the animal breed based on the physiological image.

[0009] Furthermore, the training method of the physiological indicator multi-task analysis model includes: Constructing a second training data set, wherein the images of the second training data set include physiological images of animals marked with specific abnormality types; Inputting the images in the second training data set into the physiological indicator multi-task analysis model to obtain an abnormality type recognition result; Determine the loss value based on the difference between the recognition result and the specific anomaly type marked; The physiological indicator multi-task analysis model is iteratively trained according to the loss value.

[0010] Furthermore, the step S8 of adjusting the environmental parameter values ​​of the breeding equipment according to the environmental control strategy includes the following steps: S8.1. Configure the breeding equipment with an environmental sensor to obtain the current environmental parameter values ​​of the breeding equipment; S8.2. Obtain target environmental parameter values ​​from the environmental control strategy database based on the animal species and the specific abnormality type identified; S8.3. Adjust the breeding equipment environmental parameter value based on the difference between the current breeding equipment environmental parameter value and the target environmental parameter value.

[0011] Furthermore, the method for locating the breeding area affected by the environment in S2 comprises the following steps: S2.1. Record the initial environmental parameter settings of the breeding equipment and determine the amount of single parameter adjustment; S2.2. Record the number of times the environmental parameters of the breeding equipment are adjusted, and determine the position adjustment offset of the breeding area currently affected by the environment based on the number of adjustments.

[0012] Furthermore, the step of obtaining a physiological image according to the location of the breeding area affected by the environment in S3 includes the following steps: S3.1. Determine the time point for each adjustment of the breeding equipment environmental parameters as the trigger point for the monitoring equipment to take photos; S3.2. Plan the number of image units to be shot in a single shot and select clear images as analysis images.

[0013] Furthermore, the specific abnormality type includes at least one visual pathological feature of skin lesion features, limb movement disorder features, and respiratory abnormality features; The physiological indicator multi-task analysis model configures an independent feature extraction channel for each visualized pathological feature.

[0014] Furthermore, the S7 includes the following sub-steps: S7.1. Generate a first search index based on the animal species; S7.2. Generate a second search index based on the specific exception type; S7.3. Query the target environmental parameter value in the environmental control strategy database based on the first retrieval index and the second retrieval index. In a second aspect of the present invention, a breeding management system based on controlled animal husbandry is provided, comprising: An environmental parameter adjustment module is used to adjust the environmental parameters of the breeding equipment so that the environmental parameters affect the breeding area to be managed; The breeding area positioning module is used to locate the breeding area affected by the environment through the environmental parameters of the current breeding equipment; A physiological image acquisition module, configured in a monitoring device of the breeding equipment, for acquiring physiological images according to the location of the breeding area affected by the environment; a breed identification module, configured to input the physiological image into a breed identification model and determine the animal breed in the physiological image using the breed identification model; a physiological characteristic analysis module, configured to input the physiological image into a physiological index multi-task analysis model to identify a specific abnormality type in the physiological image; An environmental control strategy database module is used to establish an environmental control strategy database, wherein the environmental control strategy database is used to store a mapping relationship between a specific abnormality type corresponding to each animal species and a target environmental parameter; An environment control strategy matching module is used to match the corresponding environment control strategy according to the specific abnormality type; The environmental parameter control module is used to adjust the environmental parameter values ​​of the breeding equipment according to the environmental regulation strategy.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. By adjusting the environmental parameters of the breeding equipment and accurately locating the affected areas, precise control of the breeding environment can be achieved, ensuring that the animals are in a suitable growth environment, thereby improving breeding efficiency and animal health.

[0016] 2. Using the breed recognition model and the physiological indicator multi-task analysis model, the animal breed can be automatically identified and the physiological characteristics can be analyzed to determine the type of abnormality, reducing manual intervention, improving recognition accuracy and efficiency, and reducing the risk of misjudgment. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily apparent by reading the following detailed description with reference to the accompanying drawings, in which several embodiments of the present invention are shown by way of example and not limitation, in which: Figure 1A schematic flow chart of a breeding management method based on controllable animal husbandry provided by one embodiment of the present invention; Figure 2 A schematic structural diagram of a breeding management system based on controllable animal husbandry provided by one embodiment of the present invention; Figure 3 The figure schematically shows the structure of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0018] The principles and spirit of the present invention will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are provided solely to enable those skilled in the art to better understand and implement the present invention, and are not intended to limit the scope of the present invention in any way. Rather, these embodiments are provided to make the present invention more thorough and complete, and to fully convey the scope of the present invention to those skilled in the art.

[0019] Those skilled in the art will appreciate that the embodiments of the present invention may be implemented as a system, apparatus, device, method, or computer program product. Therefore, the present invention may be implemented in the following forms: entirely in hardware, entirely in software (including firmware, resident software, microcode, etc.), or in a combination of hardware and software.

[0020] It should be noted that any number of elements in the drawings is for illustration only and not for limitation, and any naming is only for distinction and does not have any limiting meaning.

[0021] Reference below Figure 1 , Figure 1 This is a flow chart of a breeding management method based on controllable animal husbandry provided by one embodiment of the present invention. Figure 1 As shown, a breeding management approach based on controlled animal husbandry includes: S5. Inputting the physiological image into a physiological index multi-task analysis model to identify a specific abnormality type in the physiological image; S6. Establishing an environmental control strategy database, wherein the environmental control strategy database is used to store a mapping relationship between a specific abnormality type corresponding to each animal species and a target environmental parameter; S7. Matching a corresponding environmental control strategy according to the specific abnormality type; S8. Adjust the environmental parameter values ​​of the breeding equipment according to the environmental control strategy.

[0022] It should be noted that the breeding management method in the embodiments of the present invention first involves adjusting the environmental parameters of the breeding equipment. Environmental parameters here refer to various factors that can affect animal growth and health during the breeding process, such as temperature, humidity, and light intensity. By adjusting these parameters, a suitable growth environment is created for the animals, thereby affecting the breeding areas to be managed. Next, the current environmental parameters of the breeding equipment are used to locate the breeding areas affected by the environment. This means that the affected areas must be determined based on changes in environmental parameters. The breeding equipment is equipped with a monitoring device to capture physiological images based on the locations of the breeding areas affected by the environment. Physiological images here refer to images of the animal's physiological characteristics captured by the monitoring device, which can reflect information such as the animal's appearance and behavior. The physiological images are input into a breed recognition model, an algorithm based on machine learning or deep learning, that can identify the animal breed in the image based on the input image data. This model can determine the specific breed of the animal in the image. A multi-task analysis model for physiological indicators is used to identify specific abnormality types in the physiological images, where specific abnormality types refer to visual pathological features of the animal's body, such as skin lesions, limb movement disorders, or respiratory abnormalities. The environmental control strategy database stores mappings between specific anomaly types and target environmental parameters for different animal species. This database provides a scientific basis for subsequent adjustments to environmental parameters based on animal anomalies. Matching specific anomaly types to corresponding environmental control strategies is achieved by querying the environmental control strategy database. Finally, adjusting environmental parameter values ​​based on the environmental control strategy is the core of the entire system. This step enables precise regulation of the breeding environment to meet the health needs of the animals.

[0023] Specifically, the environmental parameter adjustment module is responsible for adjusting the environmental parameters of the breeding equipment, including but not limited to temperature, humidity, light intensity, and ventilation rate. These parameters are adjusted using corresponding sensors and controllers. For example, a temperature sensor can monitor the ambient temperature in real time, while a temperature controller adjusts the temperature based on a preset value. The breeding area location refers to the specific area within the breeding equipment that is affected by the adjustment of specific environmental parameters. For example, a large breeding shed is divided into multiple areas, each with independently adjustable environmental parameters. The physiological image acquisition module is configured with monitoring equipment at the breeding equipment. These monitoring devices can be high-definition cameras that capture physiological images of animals. The breed identification model is a complex algorithm that determines the animal's breed by analyzing detailed features in physiological images. These detailed features can include the animal's coat color, body shape, and facial features, which are stored in a breed identification database. The multi-task analysis model for physiological indicators can identify specific abnormalities in physiological images, such as visual pathological features such as skin erythema, joint swelling, or shortness of breath. The Environmental Control Strategy Database Module stores mappings between specific anomaly types and target environmental parameters for each animal species. These mappings are based on extensive veterinary knowledge and expert experience. The Anomaly Type Matching Module searches for the corresponding environmental control strategy based on the identified anomaly type. The Environmental Parameter Control Module adjusts environmental parameter values ​​based on the environmental control strategy to ensure the animals maintain an optimal growth environment.

[0024] Preferably, when adjusting environmental parameters, the environmental parameter adjustment module precisely controls them according to a preset parameter range and adjustment step size. For example, the temperature adjustment step size can be set to 0.5 degrees Celsius, and the humidity adjustment step size can be set to 5%. When determining the location of the breeding area, the breeding area positioning module combines the initial settings and the number of adjustments of the environmental parameters and calculates the current offset using a simple mathematical model. When acquiring physiological images, the physiological image acquisition module triggers the monitoring device to capture images based on the time point of environmental parameter adjustment, ensuring that the acquired images reflect the animal's state after the environmental parameter changes. The process of constructing the breed recognition model involves collecting a large number of animal physiological images as training data. These images are labeled with the corresponding animal breed labels. Through a machine learning algorithm, the model learns the relationship between these images and breed labels, enabling it to accurately identify animal breeds in new images. The training of the physiological indicator multi-task analysis model focuses on identifying specific abnormality types. It requires a dataset of animal physiological images labeled with specific abnormality types, with independent feature extraction channels configured for each visualized pathological feature. When matching environmental control strategies, the anomaly type matching module generates a first search index based on the animal species and a second search index based on the specific anomaly type. Using these two indexes, the target environmental parameter values ​​are then retrieved from the database. When adjusting environmental parameter values, the environmental parameter control module uses a feedback control algorithm based on the target parameter values ​​in the environmental control strategy database and the current environmental parameters to achieve an optimal breeding environment.

[0025] In some embodiments, the breed identification model includes a breed identification database, and the breed identification database includes detailed characteristics of each animal breed; Determining the animal species in the physiological image using the species recognition model includes: acquiring a first detail feature set in the physiological image; Comparing the first detailed feature set with the detailed features of each animal breed in the breed identification database to obtain the detailed features that overlap between the two and the number of the overlapping detailed features; Compare the number of overlapping detail features with the overlapping detail feature threshold, and rank the number of overlapping detail features corresponding to various types of animal breeds. If the number of overlapping detail features between the physiological image and a certain animal breed exceeds the overlapping detail feature threshold and the number of overlapping detail features is the largest, then output that the animal breed in the physiological image is the animal breed.

[0026] It should be noted that when the breed recognition model determines the animal breed in the physiological image, it will use the breed recognition database for comparison. The breed recognition database is a database that stores the detailed features of each animal breed. These detailed features can be the animal's coat color, body shape, facial features, etc. By comparing the detailed features in the physiological image with the features in the database, the animal breed in the image can be determined. Specifically, the breed recognition model will obtain the first set of detailed features in the physiological image, and then compare these features with the detailed features of each animal breed in the breed recognition database to obtain the detailed features that overlap between the two and the number of overlapping detailed features. If the number of overlapping detailed features between the physiological image and a certain animal breed exceeds the preset overlapping detailed feature threshold and is the largest among all breeds, the model will output the animal breed. This method can improve the accuracy of breed recognition, especially in complex breeding environments where the appearance of animals may be affected by multiple factors.

[0027] Specifically, the first detail feature set in the breed recognition model refers to a set of features extracted from the physiological image. These features may include the animal's coat color, body shape, facial features, etc. The detail features in the breed recognition database refer to various pre-stored features used to distinguish different animal breeds. These features are obtained by analyzing a large number of sample images and can accurately reflect the differences in appearance of animals of different breeds. The number of overlapping detail features refers to the number of features in the physiological image that match the characteristics of a certain breed in the database. The overlapping detail feature threshold is a preset value used to determine whether the animal in the physiological image belongs to a certain breed. If the number of overlapping detail features exceeds this threshold and is the largest among all breeds, then the breed of the animal in the physiological image can be determined. The setting of this threshold can be adjusted according to the actual breeding environment and the diversity of animal breeds to ensure the accuracy of recognition.

[0028] Preferably, the process of building a breed recognition model includes the following steps: First, a first training dataset containing a large number of animal physiological images is constructed, and these images are labeled with the corresponding animal breed labels. Then, this dataset is input into the breed recognition model for training. During the training process, the model analyzes the detailed features in the image and compares them with the breed labels to learn how to identify different breeds of animals. The model training will determine a first loss value based on the difference between the animal breed label and the breed analysis result, and then iteratively update the model training based on this loss value. In this way, the model can be continuously optimized to improve the accuracy of recognition.

[0029] In practice, the model extracts a first set of detail features from the input physiological image and compares them with features in the breed identification database. During this comparison, the model calculates the number of overlapping detail features and compares them to a threshold for overlapping detail features. If a breed has the highest number of overlapping detail features among all breeds, the model outputs that breed as the identification result. This approach can effectively improve the accuracy and efficiency of breed identification, especially in complex breeding environments where an animal's appearance may be affected by a variety of factors.

[0030] In some embodiments, the method for training the variety recognition model includes: constructing a first training data set, The first training data set includes physiological images and corresponding animal species labels; Inputting the first training data set into the breed recognition model to perform animal breed analysis, and determining a first loss value based on a difference between the animal breed label and the breed analysis result; The breed recognition model is iteratively updated and trained according to the first loss value, and the trained breed recognition model is used to determine the animal breed based on the physiological image.

[0031] It should be noted that the training method of the breed recognition model is an important part of the embodiment of the present invention. The training process involves constructing a first training data set containing physiological images and corresponding animal breed labels. The physiological images here refer to the appearance images of animals obtained by monitoring equipment, which can reflect the breed characteristics of the animals. The animal breed labels refer to the animal breed information corresponding to these images, which are used to train the model to identify animals of different breeds. By inputting the first training data set into the breed recognition model, the model can perform animal breed analysis and determine the first loss value based on the difference between the analysis results and the actual breed labels. This first loss value is an indicator that measures the difference between the model's predicted results and the actual results, and is used to evaluate the training effect of the model. According to this loss value, the breed recognition model is iteratively updated and trained, and finally a model that can accurately judge animal breeds based on physiological images is obtained.

[0032] Specifically, the construction of the first training dataset forms the foundation for training the breed recognition model. This dataset contains a large number of physiological images labeled with animal breeds, covering a variety of possible breed characteristics, such as coat color, body shape, and facial features. During training, the model analyzes the detailed features in these images and compares them with the breed labels. The first loss value here is calculated by calculating the difference between the breed predicted by the model and the actual breed label. This loss value can be a numerical value that represents the degree of mismatch between the predicted result and the actual result. The iterative update training process of the model adjusts the model parameters based on this loss value to reduce the prediction error. This process is repeated until the difference between the model's predicted result and the actual breed label reaches an acceptable range. In this process, the model learns how to extract key features from physiological images and match these features to specific animal breeds.

[0033] Preferably, the process for building a breed recognition model can be further refined. First, when constructing the first training dataset, a large number of animal physiological images must be collected, ensuring sufficient diversity and representativeness. These images can be from different angles, under different lighting conditions, and in different background environments to ensure that the model can accurately identify animal breeds in a variety of practical breeding scenarios. Each image must be accurately labeled with the corresponding animal breed label. These images and labels are then input into the breed recognition model for training. During training, the model uses a specific algorithm, such as a convolutional neural network (CNN), to analyze image features. This algorithm automatically extracts key image features such as edges, texture, and shape. By comparing these features with the breed labels, the model calculates a first loss value and adjusts its parameters based on this loss value. This process is repeated until the model achieves a satisfactory level of prediction accuracy. The resulting breed recognition model can accurately identify the animal breed in images based on the input physiological images, providing important basic information for subsequent breeding management.

[0034] In some embodiments, the training method of the multi-task analysis model of physiological indicators includes: Constructing a second training data set, wherein the images of the second training data set include physiological images of animals marked with specific abnormality types; Inputting the images in the second training data set into the physiological indicator multi-task analysis model to obtain an abnormality type recognition result; Determine the loss value based on the difference between the recognition result and the specific anomaly type marked; The physiological indicator multi-task analysis model is iteratively trained according to the loss value.

[0035] It should be noted that the training method for the multi-task analysis model for physiological indicators is a key component of the embodiments of the present invention. This method involves constructing a second training dataset specifically labeled with specific abnormality types. These images cover visual pathological features of different animal species, including typical pathological manifestations such as skin lesions, limb movement disorders, and respiratory abnormalities. Each image is precisely labeled with the corresponding specific abnormality type label, determined based on veterinary diagnostic standards, to guide the model in learning to recognize different pathological features. By inputting images from the second training dataset into the multi-task analysis model for physiological indicators, the model outputs a recognition result for the specific abnormality type in the image. A loss value is calculated based on the difference between the model's recognition result and the labeled specific abnormality type. The model parameters are iteratively optimized based on this loss value, ultimately training a specialized model that can accurately identify specific abnormality types. In terms of model architecture design, independent feature extraction channels are configured for each visual pathological feature type. For example, skin lesions are analyzed using a high-resolution texture analysis channel, limb movement disorders are analyzed using a motion trajectory tracking channel, and respiratory abnormalities are analyzed using a chest contour dynamic analysis channel. The underlying convolutional layers are shared to improve feature reuse efficiency.

[0036] Specifically, the construction of the second training dataset is the basic guarantee for model training. This dataset contains a rich and diverse sample of animal pathology images. All images are annotated by a professional veterinary team to ensure the accuracy of the labels of specific abnormality types. For example, the annotation of skin lesion features includes the area of ​​erythema, the degree of ulceration and the location of the hair loss area; the annotation of limb movement disorder features includes the degree of joint swelling and the type of gait abnormality; the annotation of respiratory abnormality features includes abnormal respiratory rate and characteristic posture changes. During the training process, the model compares the input image with these accurately annotated specific abnormality type labels, and quantifies the degree of difference between the predicted results and the actual labels through the cross-entropy loss function. This loss value serves as a guiding indicator for model optimization. The network weight parameters are continuously adjusted through the back-propagation algorithm, so that the model gradually improves the recognition accuracy of various specific abnormality types. The training process pays special attention to sample balance processing, and oversampling technology is used for rare pathological types to avoid the model being biased towards common diseases.

[0037] Preferably, a systematic implementation process should be employed to construct a multi-task analysis model for physiological indicators. First, during the dataset construction phase, pathological images are collected across various lighting conditions, animal postures, and camera angles to establish a dedicated sample library for each specific abnormality type. For example, skin lesion samples should include images from various body surface locations and stages; movement disorder samples should include video frame sequences from various motion states, such as standing and walking. Model training utilizes transfer learning techniques, using a pre-trained ResNet network as the base feature extractor and adding specialized branch networks for different pathological feature types. The training process applies a focal loss function to address sample imbalance, and data augmentation techniques such as random occlusion and brightness adjustment are employed to enhance model robustness. During the validation phase, an independent test set is established, containing gold-standard samples double-checked by veterinarians. Precision and recall metrics are calculated for each type of abnormality. Recognition results with confidence levels below a set threshold are manually reviewed to ensure the reliability of abnormality diagnoses. The resulting model can accurately identify specific pathological features, such as "secondary dorsal skin erythema" or "forelimb joint swelling with gait abnormality," providing a precise basis for abnormality diagnosis for environmental control.

[0038] In some embodiments, the step S8 adjusts the environmental parameter values ​​of the breeding equipment according to the environmental control strategy as follows: S8.1. Configure the breeding equipment with an environmental sensor to obtain the current environmental parameter values ​​of the breeding equipment; S8.2. Obtain target environmental parameter values ​​from the environmental control strategy database based on the animal species and the specific abnormality type identified; S8.3. Adjust the breeding equipment environmental parameter value based on the difference between the current breeding equipment environmental parameter value and the target environmental parameter value.

[0039] It should be noted that adjusting the environmental parameter values ​​according to the environmental control strategy is the core control link in the embodiment of the present invention. This process aims to provide customized environmental conditions for animals in different pathological states through precise control driven by abnormality types. Environmental parameter values ​​refer to the specific numerical values ​​of physical factors that can be controlled during animal breeding, including key indicators such as temperature, humidity, and light intensity. The environmental control strategy database stores a set of control rules verified by veterinary experience. The database establishes a mapping relationship between animal breeds, specific abnormality types, and target environmental parameters. For example, when the "skin heat rash" abnormality is identified, a cooling strategy corresponds to it, and when "respiratory mucosal edema" is identified, a humidification strategy corresponds to it. By obtaining the current environmental parameter values ​​and comparing them with the target values, precise environmental control based on pathological diagnosis is achieved.

[0040] Specifically, the environmental parameter adjustment process is implemented through a closed-loop control system. First, a distributed environmental sensor network collects real-time parameter values ​​of the breeding area, including temperature probes, humidity sensors, and light meters.

[0041] During the control execution phase, the system generates a composite search key-value query database based on the animal species output by the species identification module and the specific abnormality type (e.g., "secondary erythema on the back skin") identified by the physiological characteristics analysis module. After obtaining the target parameter value, the control system calculates the deviation between the current value and the target value and generates control instructions using the PID algorithm. For example, if the current temperature is 28°C and the target value is 25°C, the refrigeration equipment will be activated until the temperature difference falls below the stable threshold of 0.5°C.

[0042] Optimally, the environmental control process can be optimized with a three-level execution mechanism. The first level is real-time fine-tuning: Environmental sensors upload data every 5 seconds. When parameters deviate from the target value by ±5%, compensation mechanisms are automatically triggered (for example, a temperature deviation exceeding 1°C activates the thermostat). The second level is policy verification: After each control, monitoring equipment collects video streams of animal behavior, and a lightweight neural network is used to verify the anomaly mitigation effect (for example, the reduction rate of skin erythema area). The third level is policy optimization: a feedback learning mechanism is established. When the environmental control effectiveness of a specific anomaly type consistently falls below 85%, the mapping rule is automatically flagged and a veterinary review process is triggered. For safety control, parameter boundary protection is implemented (for example, temperature adjustment increments of no more than 5°C per hour) to prevent drastic environmental fluctuations. This mechanism enables precise and differentiated control of "heat rash cooling, frostbite warming" while continuously optimizing the effectiveness of the control strategy.

[0043] In some embodiments, the method for locating the breeding area affected by the environment in S2 comprises the following steps: S2.1. Record the initial environmental parameter settings of the breeding equipment and determine the amount of single parameter adjustment; S2.2. Record the number of times the environmental parameters of the breeding equipment are adjusted, and determine the position adjustment offset of the breeding area currently affected by the environment based on the number of adjustments.

[0044] It should be noted that locating the position of the breeding area affected by the environment is achieved by recording the initial environmental parameter settings of the breeding equipment and determining the amount of a single parameter adjustment. This method uses the historical records of environmental parameter adjustments to determine the location of the currently affected area. The initial environmental parameter settings refer to the parameter values ​​set by the equipment before starting to adjust the environmental parameters. These parameter values ​​may include temperature, humidity, etc. The amount of a single parameter adjustment refers to the specific value that changes each time the environmental parameters are adjusted, such as adjusting the temperature by 0.5 degrees Celsius or adjusting the humidity by 5% each time. By recording the number of these adjustments and the amount of each adjustment, the offset of the current breeding area affected by the environment can be calculated, thereby determining the specific location.

[0045] Specifically, the process of locating the position of the breeding area affected by the environment involves the following key concepts. The initial environmental parameter settings refer to the environmental parameter values ​​set when the breeding equipment starts running. These values ​​are the benchmark for subsequent adjustments. The single parameter adjustment amount refers to the specific value that changes each time the environmental parameters are adjusted, such as the temperature is adjusted by 0.5 degrees Celsius each time and the humidity is adjusted by 5% each time. The number of adjustments refers to the total number of times the environmental parameters have been adjusted since the initial settings. Through these data, the offset of the current position of the breeding area affected by the environment can be calculated. For example, if the initial temperature is set to 20 degrees Celsius and adjusted by 0.5 degrees Celsius each time, after 4 adjustments, the current temperature is 22 degrees Celsius, and the corresponding breeding area position will also shift accordingly. This positioning method is based on the adjustment history of environmental parameters and can accurately determine the position of the area affected by the environment.

[0046] Preferably, the process of locating the position of the breeding area affected by the environment can be further refined. First, record the initial environmental parameter settings of the breeding equipment, including the initial values ​​of parameters such as temperature, humidity, and light intensity. Then, determine the specific adjustment amount each time the environmental parameters are adjusted, such as adjusting the temperature by 0.5 degrees Celsius each time and adjusting the humidity by 5% each time. Next, record the number of adjustments each time, and use these data to calculate the offset of the current position of the breeding area affected by the environment. For example, if the initial temperature is 20 degrees Celsius and the adjustment is 0.5 degrees Celsius each time, after 4 adjustments, the current temperature is 22 degrees Celsius, and the corresponding breeding area position will also be offset accordingly. In this way, the position of the breeding area currently affected by the environment can be accurately determined, providing accurate information for subsequent monitoring and management. This method not only improves the accuracy of positioning, but also can reflect the impact of environmental parameter adjustments on the breeding area in real time, thereby better supporting breeding management decisions.

[0047] In some embodiments, the step of obtaining a physiological image according to the location of the breeding area affected by the environment in S3 includes the following steps: S3.1. Determine the time point for each adjustment of the breeding equipment environmental parameters as the trigger point for the monitoring equipment to take photos; S3.2. Plan the number of image units to be shot in a single shot and select clear images as analysis images.

[0048] It should be noted that the process of acquiring physiological images based on the location of the breeding area affected by the environment is achieved by determining the time points for each adjustment of the environmental parameters of the breeding equipment and using these time points as trigger points for the monitoring equipment to capture images. This method ensures that when environmental parameters change, physiological images of the animals can be acquired in a timely manner, so as to analyze the impact of environmental changes on the animals' physiological condition. Physiological images refer to images of the physiological characteristics of animals captured by monitoring equipment. These images can reflect information such as the animal's appearance and behavior. By planning the number of units of image capture in a single shot and selecting clear images from them as analysis images, the accuracy and efficiency of image analysis can be improved.

[0049] Specifically, the time point of each adjustment of the environmental parameters of the breeding equipment refers to the specific moment when the environmental parameters, such as temperature and humidity, are adjusted. These time points are used as trigger points for the monitoring equipment to capture physiological images, ensuring that the physiological images of the animals can be obtained in a timely manner when the environmental parameters change. The unit number of single-shot images refers to the number of images captured by the monitoring equipment each time the capture is triggered. For example, the monitoring equipment can capture 5 images each time the environmental parameters are adjusted. Clear images refer to those images that are of high quality and can clearly reflect the physiological characteristics of the animals, among these captured images, selected as analysis images. In this way, it can be ensured that the acquired images have sufficient quality and information content for subsequent breed identification and physiological characteristic analysis.

[0050] Preferably, the process of acquiring physiological images can be further refined. First, determine the specific time points for each adjustment of the environmental parameters of the breeding equipment, and these time points can be automatically recorded by the environmental parameter adjustment system. For example, when the temperature is adjusted from 20 degrees Celsius to 22 degrees Celsius, record the specific time when the adjustment occurs. Then, set these time points as trigger points for the monitoring equipment to capture, ensuring that the physiological images of the animals can be captured in time when the environmental parameters change. Next, plan the number of image units to be captured each time, such as 5 images each time, to ensure that there are enough images to choose from. Finally, through the image processing algorithm, select a clear image from the captured images as the analysis image. For example, an edge detection algorithm can be used to evaluate the clarity of the image, and the image with the clearest edge can be selected as the analysis image. In this way, it can be ensured that the acquired images can accurately reflect the physiological condition of the animal, thereby providing reliable data support for subsequent analysis.

[0051] In some embodiments, the specific abnormality type includes at least one visual pathological feature selected from the group consisting of skin lesion features, limb movement disorder features, and respiratory abnormality features; The physiological indicator multi-task analysis model configures an independent feature extraction channel for each visualized pathological feature.

[0052] Specific abnormality types are limited to visual pathological features, including three types of quantifiable and identifiable pathological manifestations: skin lesion features, limb movement disorder features, and respiratory abnormality features. Skin lesion features cover epidermal abnormalities such as erythema, ulceration, and hair loss areas on the body surface that can be determined through image analysis; limb movement disorder features include kinematic indicators such as joint swelling, abnormal gait angles, and limb weight distribution; and respiratory abnormality features involve dynamic signs such as abnormal respiratory rate and abnormal chest rise and fall. To ensure recognition accuracy, the multi-task analysis model for physiological indicators adopts a multi-channel parallel architecture, with independent feature extraction channels configured for each type of pathological feature: the skin lesion analysis channel uses a high-resolution texture segmentation algorithm, the limb movement channel integrates optical flow motion trajectory tracking, and the respiratory abnormality channel deploys a time-domain convolutional neural network to capture dynamic features. Each channel decodes specialized pathological features based on shared underlying image features, and ultimately outputs a diagnostic conclusion for the specific abnormality type through a feature fusion layer.

[0053] Specifically, during the implementation process, a standardized pathological feature library needs to be established as a recognition benchmark. For skin lesion characteristics, a standard atlas of erythema area percentages from level 0 to 3 and samples of ulcer depth grading are defined; for limb movement disorder characteristics, a database of healthy joint dimensions is collected and a gait cycle kinematic model is established; for respiratory abnormalities, a standard respiratory frequency video library is recorded and key frames of abnormal respiratory postures are marked. During model training, each independent channel adopts a targeted data enhancement strategy: the skin channel adds simulated hair occlusion and stain interference, the motion channel adds motion blur and perspective transformation, and the respiratory channel synthesizes changes in chest contours under different lighting conditions. A cross-validation mechanism is used in the verification stage to test the recognition accuracy of each channel using pathological samples of different species such as pigs, cattle, and sheep to ensure that the model has cross-species adaptability.

[0054] Optimally, a standard for annotating pathological features is first established, with a team of veterinary experts annotating three types of pathological features according to unified standards. Skin lesions are annotated using a grid method to quantify the proportion of abnormal areas, limb movement disorders are annotated using a joint angle measurement tool, and respiratory abnormalities are annotated by recording the number of chest rises and falls per unit time. A real-time feedback mechanism is implemented during model deployment. When the recognition confidence level falls below 90%, multi-frame verification mode is automatically triggered, improving diagnostic reliability by analyzing three consecutive image frames. For complex cases, a combined diagnostic process is implemented. For example, when "secondary skin erythema" and "abnormal respiratory rate" are simultaneously identified, an emergency alarm protocol is activated and a comprehensive environmental control plan is recommended.

[0055] In some embodiments, the step S7 includes the following sub-steps: S7.1. Generate a first search index based on the animal species; S7.2. Generate a second search index based on the specific exception type; S7.3. Query the target environmental parameter value in the environmental control strategy database based on the first retrieval index and the second retrieval index.

[0056] It should be noted that precise matching of environmental control strategies is achieved through a dual-index search mechanism, a core technology that ensures the scientific integrity of control plans. This mechanism comprises three standardized steps: first, a primary search index is generated based on animal species, classifying species such as cattle, pigs, and sheep according to a standardized coding system; second, a secondary search index is generated based on the specific abnormality type, annotating the abnormality type using the pathology coding system (e.g., skin erythema is coded P0102, joint swelling is coded M0203); and finally, a composite index key is used to query the environmental control strategy database for target environmental parameter values. The database uses a relational storage structure, with the primary key consisting of the species code and the pathology code. Associated fields contain target parameter values ​​such as temperature adjustment and humidity adjustment. This structured query approach avoids control conflicts such as requiring warming for frostbite and cooling for heat rash, ensuring targeted environmental interventions for different pathological conditions.

[0057] Specifically, the implementation of the dual-index retrieval mechanism requires the establishment of a standardized coding system. Species codes use the International Classification of Animals code (e.g., BOV for cattle, SUS for pigs), and specific abnormality types are coded according to veterinary pathology standards (e.g., DER01 for skin erythema, RES03 for shortness of breath). During a query, the system automatically executes the following process: first, the identified animal species is mapped to standard codes such as BOV to generate the first index. Then, diagnostic conclusions such as "secondary erythema on the back" are mapped to levels DER01-02 to generate the second index. Finally, the target environmental parameters are retrieved through SQL query statements. Database records contain detailed control parameters. For example, a typical record for "cattle + second-level skin erythema" is: {Temperature adjustment: -3±0.5℃, humidity adjustment: -5%±2%, light adjustment: -15%±3%}, and comes with a validity certification mark signed by the veterinary team.

[0058] Optimally, a query performance monitoring module should be established first to record the response time and result accuracy of each search. When the failure rate of a specific composite index query exceeds 5%, the database expansion process should be automatically triggered. Secondly, a fuzzy matching algorithm should be developed. When encountering a new pathology combination (such as "swine" + unknown skin lesions), the algorithm should match the closest known regulatory strategy based on the similarity of the pathological features (for example, matching the "swine" + heat rash level 3 strategy).

[0059] The above embodiments of the present invention have the following beneficial effects: 1. By adjusting the environmental parameters of the breeding equipment and accurately locating the affected areas, precise control of the breeding environment can be achieved, ensuring that the animals are in a suitable growth environment, thereby improving breeding efficiency and animal health.

[0060] 2. Using the breed recognition model and the physiological indicator multi-task analysis model, the animal breed can be automatically identified and its physiological characteristics can be analyzed to determine the type of abnormality, reducing manual intervention, improving recognition accuracy and efficiency, and reducing the risk of misjudgment.

[0061] like Figure 2 As shown, some embodiments of a breeding management system based on controllable animal husbandry include: Environmental parameter adjustment module 201, used to adjust the environmental parameters of the breeding equipment so that the environmental parameters affect the breeding area to be managed; The breeding area positioning module 202 is used to locate the breeding area affected by the environment through the environmental parameters of the current breeding equipment; The physiological image acquisition module 203 is a monitoring device configured in the breeding equipment, and is used to acquire physiological images according to the location of the breeding area affected by the environment; A breed identification module 204, configured to input the physiological image into a breed identification model and determine the animal breed in the physiological image using the breed identification model; A physiological characteristic analysis module 205 is configured to input the physiological image into a physiological index multi-task analysis model to identify a specific abnormality type in the physiological image; An environmental control strategy database module 206 is used to establish an environmental control strategy database, wherein the environmental control strategy database is used to store a mapping relationship between a specific abnormality type corresponding to each animal species and a target environmental parameter; An environment control strategy matching module 207 is used to match a corresponding environment control strategy according to the specific abnormality type; The environmental parameter control module 208 is used to adjust the environmental parameter values ​​of the breeding equipment according to the environmental regulation strategy.

[0062] It is understandable that the modules and references in the breeding management system based on controllable animal husbandry are Figure 1 The steps in the breeding management method based on controlled animal husbandry described above correspond to each other. Therefore, the operations, features and beneficial effects described above for the breeding management method based on controlled animal husbandry are also applicable to the breeding management system based on controlled animal husbandry and the modules contained therein, and will not be repeated here.

[0063] Reference below Figure 3, which shows a schematic structural diagram of an electronic device 300 suitable for implementing some embodiments of the present invention. The electronic devices in some embodiments of the present invention may include, but are not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 3 The terminal device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0064] like Figure 3 As shown, electronic device 300 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 301, which can perform various appropriate actions and processes based on programs stored in a read-only memory (ROM) 302 or programs loaded from a storage device 308 into a random access memory (RAM) 303. RAM 303 also stores various programs and data required for the operation of electronic device 300. Processing device 301, ROM 302, and RAM 303 are interconnected via a bus 304. An input / output (I / O) interface 305 is also connected to bus 304.

[0065] Typically, the following devices may be connected to the I / O interface 305: an input device 306 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 307 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 308 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 309. The communication device 309 may allow the electronic device 300 to communicate with other devices wirelessly or by wire to exchange data. Figure 3 The electronic device 300 is shown with various devices, but it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed instead. Figure 3 Each block shown in the figure may represent one device, or may represent multiple devices as needed.

[0066] Furthermore, the storage medium of the embodiment of the present application stores program instructions that can implement all the above methods, wherein the program instructions can be stored in the above storage medium in the form of a software product, including a number of instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or a terminal device such as a computer, server, mobile phone, or tablet.

[0067] The above descriptions merely illustrate some preferred embodiments of the present invention and the underlying technical principles. Those skilled in the art should understand that the scope of the invention encompassed by the embodiments of the present invention is not limited to technical solutions formed by specific combinations of the aforementioned technical features. It also encompasses other technical solutions formed by any combination of the aforementioned technical features or their equivalents, without departing from the aforementioned inventive concept. For example, a technical solution formed by replacing the aforementioned features with (but not limited to) technical features with similar functions disclosed in the embodiments of the present invention.

Claims

1. A breeding management method based on controllable animal husbandry, characterized in that: The following steps are involved: S1. Adjust the environmental parameters of the breeding equipment so that the environmental parameters affect the breeding area to be managed; S2. Locate the breeding area affected by the environment based on the environmental parameters of the current breeding equipment; S3. Configuring monitoring equipment for the breeding equipment to obtain physiological images based on the location of the breeding area affected by the environment; S4. Inputting the physiological image into a species recognition model, and determining the animal species in the physiological image using the species recognition model; S5. Inputting the physiological image into a physiological index multi-task analysis model to identify a specific abnormality type in the physiological image; S6. Establishing an environmental control strategy database, wherein the environmental control strategy database is used to store a mapping relationship between a specific abnormality type corresponding to each animal species and a target environmental parameter; S7. Matching a corresponding environmental control strategy according to the specific abnormality type; S8. Adjust the environmental parameter values ​​of the breeding equipment according to the environmental control strategy.

2. The breeding management method based on controllable animal husbandry according to claim 1, characterized in that: The breed identification model includes a breed identification database, and the breed identification database includes detailed characteristics of each animal breed; Determining the animal species in the physiological image by using the species recognition model in S4 includes: Acquiring a first detail feature set in the physiological image; Comparing the first detailed feature set with the detailed features of each animal breed in the breed identification database to obtain the detailed features that overlap between the two and the number of the overlapping detailed features; Compare the number of overlapping detail features with the overlapping detail feature threshold, and rank the number of overlapping detail features corresponding to various types of animal breeds. If the number of overlapping detail features between the physiological image and a certain animal breed exceeds the overlapping detail feature threshold and the number of overlapping detail features is the largest, then output that the animal breed in the physiological image is the animal breed.

3. The breeding management method based on controllable animal husbandry according to claim 1 or 2, characterized in that: The training method of the variety recognition model includes: Construct the first training dataset, The first training data set includes physiological images and corresponding animal species labels; Inputting the first training data set into the breed recognition model to perform animal breed analysis, and determining a first loss value based on a difference between the animal breed label and the breed analysis result; The breed recognition model is iteratively updated and trained according to the first loss value, and the trained breed recognition model is used to determine the animal breed based on the physiological image.

4. The breeding management method based on controllable animal husbandry according to claim 1, characterized in that: The training method of the physiological indicator multi-task analysis model includes: Constructing a second training data set, wherein the images of the second training data set include physiological images of animals marked with specific abnormality types; Inputting the images in the second training data set into the physiological indicator multi-task analysis model to obtain an abnormality type recognition result; Determine the loss value based on the difference between the recognition result and the specific anomaly type marked; The physiological indicator multi-task analysis model is iteratively trained according to the loss value.

5. The breeding management method based on controllable animal husbandry according to claim 1, characterized in that: Said step S8 of adjusting the environmental parameter values ​​of the breeding equipment according to the environmental control strategy comprises the following steps: S8.

1. Configure the breeding equipment with an environmental sensor to obtain the current environmental parameter values ​​of the breeding equipment; S8.

2. Obtain target environmental parameter values ​​from the environmental control strategy database based on the animal species and the specific abnormality type identified; S8.

3. Adjust the breeding equipment environmental parameter value based on the difference between the current breeding equipment environmental parameter value and the target environmental parameter value.

6. The breeding management method based on controllable animal husbandry according to claim 1, characterized in that: The method for locating the breeding area affected by the environment in S2 comprises the following steps: S2.

1. Record the initial environmental parameter settings of the breeding equipment and determine the amount of single parameter adjustment; S2.

2. Record the number of times the environmental parameters of the breeding equipment are adjusted, and determine the position adjustment offset of the breeding area currently affected by the environment based on the number of adjustments.

7. The breeding management method based on controllable animal husbandry according to claim 1, characterized in that: Acquiring a physiological image according to the location of the breeding area affected by the environment in S3 includes the following steps: S3.

1. Determine the time point for each adjustment of the breeding equipment environmental parameters as the trigger point for the monitoring equipment to take photos; S3.

2. Plan the number of image units to be shot in a single shot and select clear images as analysis images.

8. The breeding management method based on controllable animal husbandry according to claim 1, characterized in that: The specific abnormality type includes at least one visual pathological feature selected from skin lesion features, limb movement disorder features, and respiratory abnormality features; The physiological indicator multi-task analysis model configures an independent feature extraction channel for each visualized pathological feature.

9. The breeding management method based on controllable animal husbandry according to claim 1, characterized in that: The step S7 includes the following steps: matching the corresponding environment control strategy according to the specific abnormality type: S7.

1. Generate a first search index based on the animal species; S7.

2. Generate a second search index based on the specific exception type; S7.

3. Query the target environmental parameter value in the environmental control strategy database based on the first retrieval index and the second retrieval index.

10. A breeding management system based on controllable animal husbandry, characterized in that: Includes the following modules: An environmental parameter adjustment module is used to adjust the environmental parameters of the breeding equipment so that the environmental parameters affect the breeding area to be managed; The breeding area positioning module is used to locate the breeding area affected by the environment through the environmental parameters of the current breeding equipment; A physiological image acquisition module, configured in a monitoring device of the breeding equipment, for acquiring physiological images according to the location of the breeding area affected by the environment; a breed identification module, configured to input the physiological image into a breed identification model and determine the animal breed in the physiological image using the breed identification model; a physiological characteristic analysis module, configured to input the physiological image into a physiological index multi-task analysis model to identify a specific abnormality type in the physiological image; An environmental control strategy database module is used to establish an environmental control strategy database, wherein the environmental control strategy database is used to store a mapping relationship between a specific abnormality type corresponding to each animal species and a target environmental parameter; An environment control strategy matching module is used to match the corresponding environment control strategy according to the specific abnormality type; The environmental parameter control module is used to adjust the environmental parameter values ​​of the breeding equipment according to the environmental regulation strategy.

Images