A method and system for identifying and tracking cashmere goats based on a large model.

Abstract

This invention provides a method and system for identifying and tracking cashmere goats based on a large model. The method includes: collecting goat face data; using the goat face data to fine-tune and train a large model to obtain a first goat face recognition module; inputting monitoring video data of target cashmere goats in the breeding area into the first goat face recognition module; the first goat face recognition module outputting a first goat face recognition result and a second goat face recognition result; wherein the first goat face recognition result is derived based on the analysis of the static goat face data, and the second goat face recognition result is derived based on the analysis of the dynamic goat face data; fusing the first and second goat face recognition results to obtain a final goat face recognition result; and tracking the target cashmere goat based on the goat face recognition result. This invention simultaneously identifies cashmere goats based on both static and dynamic facial information, significantly improving the accuracy of cashmere goat identification and tracking.

Description

Technical Field

[0001] This invention relates to the field of smart farming technology, and more specifically, to a method and system for identifying and tracking cashmere goats based on a large model. Background Technology

[0002] As the livestock industry rapidly develops towards modernization, digitalization, and intelligence, building farms integrated with big data has become an industry trend. By collecting individual information about livestock, precision farming aims to improve efficiency, reduce costs, and achieve healthy farming practices. In this context, accurate identification of each animal becomes a prerequisite for collecting individual information.

[0003] Currently, the primary method for identifying livestock on farms is by using ear tags. However, this method suffers from low identification efficiency, high labor intensity, and the need for frequent maintenance. Furthermore, ear tags can cause discomfort and pain to livestock, and even lead to complications affecting ear integrity. Therefore, finding an efficient and non-invasive method for livestock identification is urgently needed.

[0004] In recent years, with the development of deep learning technology, scholars have begun to use these technologies to learn the biometrics of livestock in order to achieve accurate identification of livestock. Among various livestock biometric identification methods, facial recognition is considered one of the most promising methods because it does not require contact with livestock, achieving efficient identification while effectively avoiding stressful behaviors in livestock.

[0005] However, current sheep face recognition technology mainly relies on the static features of sheep faces (such as facial feature distribution, coat color, and facial contour features). In actual recognition processes, sheep faces are often in motion, and relying solely on static features can easily lead to recognition failures or errors. Therefore, further improvements to sheep face recognition methods are needed to enhance accuracy. Summary of the Invention

[0006] To address the technical problems existing in the background art, the present invention provides a cashmere goat identification and tracking method, system, electronic device, computer storage medium, and computer program product based on a large model.

[0007] This invention provides a method for identifying and tracking cashmere goats based on a large model, the method comprising the following steps:

[0008] Collect sheep face data, and use the sheep face data to fine-tune and train a large model to obtain a first sheep face recognition module; wherein, the sheep face data includes static sheep face data and dynamic sheep face data;

[0009] The monitoring video data of the target cashmere goats in the breeding area is input into the first sheep face recognition module, and the first sheep face recognition module outputs a first sheep face recognition result and a second sheep face recognition result; wherein, the first sheep face recognition result is obtained based on the analysis of the static sheep face data, and the second sheep face recognition result is obtained based on the analysis of the dynamic sheep face data;

[0010] The sheep face recognition result is obtained by fusing the first sheep face recognition result and the second sheep face recognition result;

[0011] The target cashmere goat is tracked based on the sheep face recognition results.

[0012] Optionally, the step of collecting sheep face data and using the sheep face data to fine-tune and train a large model to obtain a first sheep face recognition module includes:

[0013] When each cashmere goat in the breeding area is in a static state, its face is photographed from multiple angles and under multiple lighting conditions, and the static face data is extracted from the face images.

[0014] When each cashmere goat in the breeding area is in a facial activity state, its facial movement video is filmed, and the dynamic goat face data is extracted from the facial movement video;

[0015] The static sheep face data is used to perform preliminary fine-tuning on the large model to learn the basic features of sheep faces; and the dynamic sheep face data is used to further fine-tune the large model after preliminary fine-tuning to enhance the large model's ability to recognize the dynamic changes in sheep face features.

[0016] Optionally, before inputting the monitoring video data of the target cashmere goats within the breeding area into the first goat face recognition module, the method further includes:

[0017] Obtain feeding plan data for the breeding area, and predict the probability of cashmere goats being in a static facial state based on the feeding plan data and the current time period; wherein, the feeding plan data includes feeding time;

[0018] If the probability is lower than the probability threshold, the monitoring video data of the target cashmere goats in the breeding area will be input into the first sheep face recognition module.

[0019] If the probability is higher than the probability threshold, the monitoring video data of the target cashmere goats in the breeding area is input into the second sheep face recognition module; wherein, the second sheep face recognition module includes a convolutional feature extraction model and a sheep face recognition model, and the sheep face recognition model is constructed based on a classification algorithm.

[0020] Optionally, the monitoring video data of the target cashmere goats in the breeding area is input into the first sheep face recognition module, and the first sheep face recognition module outputs a first sheep face recognition result and a second sheep face recognition result, including:

[0021] Image sequences containing only the facial region of the target cashmere goats are extracted from the monitoring video data of the target cashmere goats in the breeding area, and the image sequences are input into the first sheep face recognition module;

[0022] The first sheep face recognition module extracts the static sheep face data and the dynamic sheep face data from the image sequence, and analyzes the static sheep face data to obtain the first sheep face recognition result.

[0023] The second sheep face recognition result is then derived based on the first sheep face recognition result and the dynamic sheep face data analysis.

[0024] Optionally, the step of fusing the first sheep face recognition result and the second sheep face recognition result to obtain the sheep face recognition result includes:

[0025] The weighting weights of the first sheep face recognition result and the second sheep face recognition result are determined respectively; wherein, the weighting weight of the first sheep face recognition result is lower than that of the second sheep face recognition result;

[0026] The sheep face recognition result is obtained by fusing the first sheep face recognition result and the second sheep face recognition result based on the weighted weights.

[0027] Optionally, tracking the target cashmere goat based on the sheep face recognition result includes:

[0028] Obtain tracking project information, and add identity tags and additional tags to the target cashmere goat based on the tracking project information.

[0029] The present invention also provides a cashmere goat identification and tracking system based on a large model. The system includes a processor and a memory, wherein the memory stores executable program code; the processor calls the executable program code stored in the memory to perform the following steps:

[0030] Collect sheep face data, and use the sheep face data to fine-tune and train a large model to obtain a first sheep face recognition module; wherein, the sheep face data includes static sheep face data and dynamic sheep face data;

[0031] The monitoring video data of the target cashmere goats in the breeding area is input into the first sheep face recognition module, and the first sheep face recognition module outputs a first sheep face recognition result and a second sheep face recognition result; wherein, the first sheep face recognition result is obtained based on the analysis of the static sheep face data, and the second sheep face recognition result is obtained based on the analysis of the dynamic sheep face data;

[0032] The sheep face recognition result is obtained by fusing the first sheep face recognition result and the second sheep face recognition result;

[0033] The target cashmere goat is tracked based on the sheep face recognition results.

[0034] Optionally, the step of collecting sheep face data and using the sheep face data to fine-tune and train a large model to obtain a first sheep face recognition module includes:

[0035] When each cashmere goat in the breeding area is in a static state, its face is photographed from multiple angles and under multiple lighting conditions, and the static face data is extracted from the face images.

[0036] When each cashmere goat in the breeding area is in a facial activity state, its facial movement video is filmed, and the dynamic goat face data is extracted from the facial movement video;

[0037] The static sheep face data is used to perform preliminary fine-tuning on the large model to learn the basic features of sheep faces; and the dynamic sheep face data is used to further fine-tune the large model after preliminary fine-tuning to enhance the large model's ability to recognize the dynamic changes in sheep face features.

[0038] Optionally, before inputting the monitoring video data of the target cashmere goats within the breeding area into the first goat face recognition module, the method further includes:

[0039] Obtain feeding plan data for the breeding area, and predict the probability of cashmere goats being in a static facial state based on the feeding plan data and the current time period; wherein, the feeding plan data includes feeding time;

[0040] If the probability is lower than the probability threshold, the monitoring video data of the target cashmere goats in the breeding area will be input into the first sheep face recognition module.

[0041] If the probability is higher than the probability threshold, the monitoring video data of the target cashmere goats in the breeding area is input into the second sheep face recognition module; wherein, the second sheep face recognition module includes a convolutional feature extraction model and a sheep face recognition model, and the sheep face recognition model is constructed based on a classification algorithm.

[0042] Optionally, the monitoring video data of the target cashmere goats in the breeding area is input into the first sheep face recognition module, and the first sheep face recognition module outputs a first sheep face recognition result and a second sheep face recognition result, including:

[0043] Image sequences containing only the facial region of the target cashmere goats are extracted from the monitoring video data of the target cashmere goats in the breeding area, and the image sequences are input into the first sheep face recognition module;

[0044] The first sheep face recognition module extracts the static sheep face data and the dynamic sheep face data from the image sequence, and analyzes the static sheep face data to obtain the first sheep face recognition result.

[0045] The second sheep face recognition result is then derived based on the first sheep face recognition result and the dynamic sheep face data analysis.

[0046] Optionally, the step of fusing the first sheep face recognition result and the second sheep face recognition result to obtain the sheep face recognition result includes:

[0047] The weighting weights of the first sheep face recognition result and the second sheep face recognition result are determined respectively; wherein, the weighting weight of the first sheep face recognition result is lower than that of the second sheep face recognition result;

[0048] The sheep face recognition result is obtained by fusing the first sheep face recognition result and the second sheep face recognition result based on the weighted weights.

[0049] Optionally, tracking the target cashmere goat based on the sheep face recognition result includes:

[0050] Obtain tracking project information, and add identity tags and additional tags to the target cashmere goat based on the tracking project information.

[0051] The present invention also provides an electronic device, comprising: a memory storing executable program code; a processor coupled to the memory; the processor calling the executable program code stored in the memory to execute the method as described in any of the preceding claims.

[0052] The present invention also provides a computer storage medium storing a computer program that, when executed by a processor, performs the method described in any of the preceding claims.

[0053] The present invention also provides a computer program product comprising a computer program stored in a computer storage medium, wherein the computer program, when executed by a processor of an electronic device, implements the method described in any of the preceding claims.

[0054] The present invention identifies cashmere goats by simultaneously using both static and dynamic information about their faces, which can significantly improve the accuracy of cashmere goat identification and, consequently, improve the accuracy of a series of tracking and processing applications based on the identification information. Attached Figure Description

[0055] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0056] Figure 1 This is a schematic diagram of a cashmere goat identification and tracking method based on a large model disclosed in an embodiment of the present invention.

[0057] Figure 2 This is a schematic diagram of a cashmere goat identification and tracking system based on a large model disclosed in an embodiment of the present invention. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0059] Please see Figure 1 This invention discloses a method for identifying and tracking cashmere goats based on a large model, the method comprising the following steps:

[0060] Collect sheep face data, and use the sheep face data to fine-tune and train a large model to obtain a first sheep face recognition module; wherein, the sheep face data includes static sheep face data and dynamic sheep face data;

[0061] The monitoring video data of the target cashmere goats in the breeding area is input into the first sheep face recognition module, and the first sheep face recognition module outputs a first sheep face recognition result and a second sheep face recognition result; wherein, the first sheep face recognition result is obtained based on the analysis of the static sheep face data, and the second sheep face recognition result is obtained based on the analysis of the dynamic sheep face data;

[0062] The sheep face recognition result is obtained by fusing the first sheep face recognition result and the second sheep face recognition result;

[0063] The target cashmere goat is tracked based on the sheep face recognition results.

[0064] In the above scheme, the present invention first collects static and dynamic face data of cashmere goats to construct training data. This training data is then used to fine-tune a general-purpose model to obtain a first face recognition module. The general-purpose model is specified as OpenAI's GPT series models, Google's BERT model, Microsoft's Azure AI platform, etc. Next, surveillance cameras are deployed in the breeding area. Surveillance video data corresponding to the target cashmere goat to be identified and tracked is retrieved from the surveillance video captured by these cameras. The first face recognition module analyzes this data in two parts: a first face recognition result based on the static face data of the target cashmere goat, and a second face recognition result based on the dynamic face data of the target cashmere goat. Then, a fusion algorithm is used to fuse the first and second face recognition results to obtain the final face recognition result. Finally, the target cashmere goat can be tracked based on its face recognition result. This tracking can be used for various application scenarios such as farm management, disease monitoring, and individual tracking.

[0065] The present invention identifies cashmere goats by simultaneously using both static and dynamic information about their faces, which can significantly improve the accuracy of cashmere goat identification and, consequently, improve the accuracy of a series of tracking and processing applications based on the identification information.

[0066] Optionally, the step of collecting sheep face data and using the sheep face data to fine-tune and train a large model to obtain a first sheep face recognition module includes:

[0067] When each cashmere goat in the breeding area is in a static state, its face is photographed from multiple angles and under multiple lighting conditions, and the static face data is extracted from the face images.

[0068] When each cashmere goat in the breeding area is in a facial activity state, its facial movement video is filmed, and the dynamic goat face data is extracted from the facial movement video;

[0069] The static sheep face data is used to perform preliminary fine-tuning on the large model to learn the basic features of sheep faces; and the dynamic sheep face data is used to further fine-tune the large model after preliminary fine-tuning to enhance the large model's ability to recognize the dynamic changes in sheep face features.

[0070] In this embodiment, monitoring cameras deployed within the breeding area can distinguish whether the faces of each cashmere goat are in a static or active state. The active state is defined as the cashmere goat being in a state such as opening its mouth, yawning, eating, or ruminating. Furthermore, multi-angle and multi-light condition images of the cashmere goat's face are captured, along with video footage of its facial movements when its face is in an active state. Static and dynamic face data are then extracted from these images.

[0071] Meanwhile, when fine-tuning the large model, static sheep face data is first used for preliminary fine-tuning training so that the large model can learn and master the basic features of sheep faces (such as facial feature distribution, coat color, facial contour features, etc.). Then, dynamic sheep face data is used to further fine-tune the large model. Thus, the large model can accurately identify the basic features of sheep faces and enhance its ability to recognize the dynamic changes of sheep faces.

[0072] Optionally, before inputting the monitoring video data of the target cashmere goats within the breeding area into the first goat face recognition module, the method further includes:

[0073] Obtain feeding plan data for the breeding area, and predict the probability of cashmere goats being in a static facial state based on the feeding plan data and the current time period; wherein, the feeding plan data includes feeding time;

[0074] If the probability is lower than the probability threshold, the monitoring video data of the target cashmere goats in the breeding area will be input into the first sheep face recognition module.

[0075] If the probability is higher than the probability threshold, the monitoring video data of the target cashmere goats in the breeding area is input into the second sheep face recognition module; wherein, the second sheep face recognition module includes a convolutional feature extraction model and a sheep face recognition model, and the sheep face recognition model is constructed based on a classification algorithm.

[0076] In this embodiment, the present invention constructs two sheep face recognition modules: a first sheep face recognition module and a second sheep face recognition module. The first sheep face recognition module is based on a large model, while the second sheep face recognition module is based on a conventional classification algorithm. Compared with the first sheep face recognition module, the second sheep face recognition module has a smaller model size and a faster actual analysis speed, but its accuracy is slightly lower. To address this, the present invention first predicts the probability of cashmere goats being in a static facial state based on the feeding plan data of the breeding area and the current time period. For example, if there is a long distance between the current time period and the feeding time, it indicates that the probability of cashmere goats feeding at the current time is low, and therefore the probability of their mouths being active is low, meaning the probability of them being in a static facial state is low. Conversely, if there is a short distance between the current time period and the feeding time, the cashmere goats are likely feeding, and their mouths are likely active, meaning the probability of them being in a static facial state is high.

[0077] Therefore, when the probability of a cashmere goat's face being stationary is below a certain threshold, the captured video data of the target cashmere goat is directly input into the first face recognition module. This module then identifies the goat based on both static and dynamic face data. Conversely, when the probability of a cashmere goat's face being stationary is above the threshold, a second face recognition module analyzes the video data of the target cashmere goat within the breeding area. This analysis is based solely on static face data. By selecting this recognition strategy, the recognition speed can be appropriately improved while ensuring accuracy.

[0078] The convolutional feature extraction model first extracts static sheep face data from the surveillance video data of the target cashmere goat. Then, the sheep face recognition model classifies the extracted static sheep face data to determine the identity of the target cashmere goat, i.e., the sheep face recognition result. The aforementioned classification algorithms include, but are not limited to, support vector machines, decision trees, and neural network algorithms.

[0079] Optionally, the monitoring video data of the target cashmere goats in the breeding area is input into the first sheep face recognition module, and the first sheep face recognition module outputs a first sheep face recognition result and a second sheep face recognition result, including:

[0080] Image sequences containing only the facial region of the target cashmere goats are extracted from the monitoring video data of the target cashmere goats in the breeding area, and the image sequences are input into the first sheep face recognition module;

[0081] The first sheep face recognition module extracts the static sheep face data and the dynamic sheep face data from the image sequence, and analyzes the static sheep face data to obtain the first sheep face recognition result.

[0082] The second sheep face recognition result is then derived based on the first sheep face recognition result and the dynamic sheep face data analysis.

[0083] In this embodiment, the first sheep face recognition module should not directly analyze and process the monitoring video data of the target cashmere goat, because the monitoring video of the target cashmere goat contains its non-facial features, as well as facial and non-facial features of other cashmere goats, which can interfere with the identification of the target cashmere goat. Therefore, this invention first extracts an image sequence containing only the facial region of the target cashmere goat from the monitoring video data, and then the first sheep face recognition module performs recognition processing on the image sequence.

[0084] The first sheep face recognition module first extracts static and dynamic sheep face data of the target cashmere goat from the image sequence, and then analyzes the static sheep face data to obtain the first sheep face recognition result. Next, based on the first sheep face recognition result and the dynamic sheep face data, a second sheep face recognition result is obtained. This approach, which simultaneously uses both the first and dynamic sheep face recognition results for secondary identification of the target cashmere goat, significantly improves the accuracy of the extracted second sheep face recognition result compared to identification based solely on dynamic sheep face data. It utilizes the first sheep face recognition result derived from the static sheep face data as a reference.

[0085] Optionally, the step of fusing the first sheep face recognition result and the second sheep face recognition result to obtain the sheep face recognition result includes:

[0086] The weighting weights of the first sheep face recognition result and the second sheep face recognition result are determined respectively; wherein, the weighting weight of the first sheep face recognition result is lower than that of the second sheep face recognition result;

[0087] The sheep face recognition result is obtained by fusing the first sheep face recognition result and the second sheep face recognition result based on the weighted weights.

[0088] In this embodiment, since the target cashmere goat's face is in motion, the second sheep face recognition result is essentially derived from the analysis of static and dynamic sheep face data. Theoretically, its confidence level should be higher than that of the first sheep face recognition result. Therefore, this invention sets the weighting weight of the first sheep face recognition result to be lower than that of the second sheep face recognition result. For example, the weighting weight of the first sheep face recognition result is set to 0.4, and the weighting weight of the second sheep face recognition result is set to 0.6.

[0089] Optionally, tracking the target cashmere goat based on the sheep face recognition result includes:

[0090] Obtain tracking project information, and add identity tags and additional tags to the target cashmere goat based on the tracking project information.

[0091] In this embodiment, tracking information for the target cashmere goat can be specified according to different needs, such as farm management, disease monitoring, and individual tracking. Therefore, after identifying the goat, an identification tag is attached to it. Simultaneously, an additional tag is determined based on the tracking information for the target cashmere goat; for example, if the tracking information is farm management, the tag can be green, and if the farm management information is disease monitoring, the tag is red. The above is merely an example and is not intended to limit the scope of protection of this invention.

[0092] See Figure 2As shown in the figure, this embodiment of the invention also provides a cashmere goat identification and tracking system based on a large model. The system includes a processor and a memory, the memory storing executable program code; the processor calls the executable program code stored in the memory to implement the following steps:

[0093] Collect sheep face data, and use the sheep face data to fine-tune and train a large model to obtain a first sheep face recognition module; wherein, the sheep face data includes static sheep face data and dynamic sheep face data;

[0094] The monitoring video data of the target cashmere goats in the breeding area is input into the first sheep face recognition module, and the first sheep face recognition module outputs a first sheep face recognition result and a second sheep face recognition result; wherein, the first sheep face recognition result is obtained based on the analysis of the static sheep face data, and the second sheep face recognition result is obtained based on the analysis of the dynamic sheep face data;

[0095] The sheep face recognition result is obtained by fusing the first sheep face recognition result and the second sheep face recognition result;

[0096] The target cashmere goat is tracked based on the sheep face recognition results.

[0097] Optionally, the step of collecting sheep face data and using the sheep face data to fine-tune and train a large model to obtain a first sheep face recognition module includes:

[0098] When each cashmere goat in the breeding area is in a static state, its face is photographed from multiple angles and under multiple lighting conditions, and the static face data is extracted from the face images.

[0099] When each cashmere goat in the breeding area is in a facial activity state, its facial movement video is filmed, and the dynamic goat face data is extracted from the facial movement video;

[0100] The static sheep face data is used to perform preliminary fine-tuning on the large model to learn the basic features of sheep faces; and the dynamic sheep face data is used to further fine-tune the large model after preliminary fine-tuning to enhance the large model's ability to recognize the dynamic changes in sheep face features.

[0101] Optionally, before inputting the monitoring video data of the target cashmere goats within the breeding area into the first goat face recognition module, the method further includes:

[0102] Obtain feeding plan data for the breeding area, and predict the probability of cashmere goats being in a static facial state based on the feeding plan data and the current time period; wherein, the feeding plan data includes feeding time;

[0103] If the probability is lower than the probability threshold, the monitoring video data of the target cashmere goats in the breeding area will be input into the first sheep face recognition module.

[0104] If the probability is higher than the probability threshold, the monitoring video data of the target cashmere goats in the breeding area is input into the second sheep face recognition module; wherein, the second sheep face recognition module includes a convolutional feature extraction model and a sheep face recognition model, and the sheep face recognition model is constructed based on a classification algorithm.

[0105] Optionally, the monitoring video data of the target cashmere goats in the breeding area is input into the first sheep face recognition module, and the first sheep face recognition module outputs a first sheep face recognition result and a second sheep face recognition result, including:

[0106] Image sequences containing only the facial region of the target cashmere goats are extracted from the monitoring video data of the target cashmere goats in the breeding area, and the image sequences are input into the first sheep face recognition module;

[0107] The first sheep face recognition module extracts the static sheep face data and the dynamic sheep face data from the image sequence, and analyzes the static sheep face data to obtain the first sheep face recognition result.

[0108] The second sheep face recognition result is then derived based on the first sheep face recognition result and the dynamic sheep face data analysis.

[0109] Optionally, the step of fusing the first sheep face recognition result and the second sheep face recognition result to obtain the sheep face recognition result includes:

[0110] The weighting weights of the first sheep face recognition result and the second sheep face recognition result are determined respectively; wherein, the weighting weight of the first sheep face recognition result is lower than that of the second sheep face recognition result;

[0111] The sheep face recognition result is obtained by fusing the first sheep face recognition result and the second sheep face recognition result based on the weighted weights.

[0112] Optionally, tracking the target cashmere goat based on the sheep face recognition result includes:

[0113] Obtain tracking project information, and add identity tags and additional tags to the target cashmere goat based on the tracking project information.

[0114] This invention also provides an electronic device, including: a memory storing executable program code; a processor coupled to the memory; the processor calling the executable program code stored in the memory to execute the method as described in any of the preceding embodiments.

[0115] This invention also provides a computer storage medium storing a computer program, which is executed by a processor to perform the method described in any of the preceding embodiments.

[0116] This invention also provides a computer program product comprising a computer program stored in a computer storage medium, wherein the computer program, when executed by a processor of an electronic device, implements the method described in any of the preceding embodiments.

[0117] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (devices), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0118] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0119] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0120] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for identifying and tracking cashmere goats based on a large model, characterized in that: The method includes the following steps: Collect sheep face data, and use the sheep face data to fine-tune and train a large model to obtain a first sheep face recognition module; wherein, the sheep face data includes static sheep face data and dynamic sheep face data; The monitoring video data of the target cashmere goats in the breeding area is input into the first sheep face recognition module, and the first sheep face recognition module outputs a first sheep face recognition result and a second sheep face recognition result; wherein, the first sheep face recognition result is obtained based on the analysis of the static sheep face data, and the second sheep face recognition result is obtained based on the analysis of the dynamic sheep face data; The sheep face recognition result is obtained by fusing the first sheep face recognition result and the second sheep face recognition result; The target cashmere goat is tracked based on the sheep face recognition results; The process of collecting sheep face data and using the sheep face data to fine-tune and train a large model to obtain a first sheep face recognition module includes: When each cashmere goat in the breeding area is in a static state, its face is photographed from multiple angles and under multiple lighting conditions, and the static face data is extracted from the face images. When each cashmere goat in the breeding area is in a facial activity state, its facial movement video is filmed, and the dynamic goat face data is extracted from the facial movement video; The static sheep face data is used to perform preliminary fine-tuning on the large model in order to learn the basic features of sheep faces; and the dynamic sheep face data is used to perform further fine-tuning on the large model after preliminary fine-tuning in order to enhance the large model's ability to recognize the dynamic changes in sheep face features. The monitoring video data of the target cashmere goats within the breeding area is input into the first sheep face recognition module. The first sheep face recognition module outputs a first sheep face recognition result and a second sheep face recognition result, including: Image sequences containing only the facial region of the target cashmere goats are extracted from the monitoring video data of the target cashmere goats in the breeding area, and the image sequences are input into the first goat face recognition module; The first sheep face recognition module extracts the static sheep face data and the dynamic sheep face data from the image sequence, and analyzes the static sheep face data to obtain the first sheep face recognition result. The second sheep face recognition result is then derived based on the first sheep face recognition result and the dynamic sheep face data analysis.

2. The cashmere goat identification and tracking method based on a large model according to claim 1, characterized in that: Before inputting the monitoring video data of the target cashmere goats within the breeding area into the first goat face recognition module, the method further includes: Obtain feeding plan data for the breeding area, and predict the probability of cashmere goats being in a static facial state based on the feeding plan data and the current time period; wherein, the feeding plan data includes feeding time; If the probability is lower than the probability threshold, the monitoring video data of the target cashmere goats in the breeding area will be input into the first sheep face recognition module. If the probability is higher than the probability threshold, the monitoring video data of the target cashmere goats in the breeding area is input into the second sheep face recognition module; wherein, the second sheep face recognition module includes a convolutional feature extraction model and a sheep face recognition model, and the sheep face recognition model is constructed based on a classification algorithm.

3. The cashmere goat identification and tracking method based on a large model according to claim 1, characterized in that: The sheep face recognition result is obtained by fusing the first sheep face recognition result and the second sheep face recognition result, including: The weighting weights of the first sheep face recognition result and the second sheep face recognition result are determined respectively; wherein, the weighting weight of the first sheep face recognition result is lower than that of the second sheep face recognition result; The sheep face recognition result is obtained by fusing the first sheep face recognition result and the second sheep face recognition result based on the weighted weights.

4. The cashmere goat identification and tracking method based on a large model according to claim 3, characterized in that: Tracking the target cashmere goat based on the sheep face recognition results includes: Obtain tracking project information, and add identity tags and additional tags to the target cashmere goat based on the tracking project information.

5. A cashmere goat identification and tracking system based on a large model, the system comprising a processor and a memory, the memory storing executable program code; characterized in that: The processor invokes the executable program code stored in the memory to perform the following steps: Collect sheep face data, and use the sheep face data to fine-tune and train a large model to obtain a first sheep face recognition module; wherein, the sheep face data includes static sheep face data and dynamic sheep face data; The monitoring video data of the target cashmere goats in the breeding area is input into the first sheep face recognition module, and the first sheep face recognition module outputs a first sheep face recognition result and a second sheep face recognition result; wherein, the first sheep face recognition result is obtained based on the analysis of the static sheep face data, and the second sheep face recognition result is obtained based on the analysis of the dynamic sheep face data; The sheep face recognition result is obtained by fusing the first sheep face recognition result and the second sheep face recognition result; The target cashmere goat is tracked based on the sheep face recognition results; The process of collecting sheep face data and using the sheep face data to fine-tune and train a large model to obtain a first sheep face recognition module includes: When each cashmere goat in the breeding area is in a static state, its face is photographed from multiple angles and under multiple lighting conditions, and the static face data is extracted from the face images. When each cashmere goat in the breeding area is in a facial activity state, its facial movement video is filmed, and the dynamic goat face data is extracted from the facial movement video; The static sheep face data is used to perform preliminary fine-tuning on the large model in order to learn the basic features of sheep faces; and the dynamic sheep face data is used to perform further fine-tuning on the large model after preliminary fine-tuning in order to enhance the large model's ability to recognize the dynamic changes in sheep face features. The monitoring video data of the target cashmere goats within the breeding area is input into the first sheep face recognition module. The first sheep face recognition module outputs a first sheep face recognition result and a second sheep face recognition result, including: Image sequences containing only the facial region of the target cashmere goats are extracted from the monitoring video data of the target cashmere goats in the breeding area, and the image sequences are input into the first goat face recognition module; The first sheep face recognition module extracts the static sheep face data and the dynamic sheep face data from the image sequence, and analyzes the static sheep face data to obtain the first sheep face recognition result. The second sheep face recognition result is then derived based on the first sheep face recognition result and the dynamic sheep face data analysis.

6. An electronic device, comprising: Memory containing executable program code; A processor coupled to the memory; characterized in that: the processor calls the executable program code stored in the memory to perform the method as described in any one of claims 1-4.

7. A computer storage medium storing a computer program, characterized in that: The computer program is executed by the processor to perform the method as described in any one of claims 1-4.

8. A computer program product, the computer program product comprising a computer program stored in a computer storage medium, characterized in that: When the computer program is executed by the processor of the electronic device, it implements the method as described in any one of claims 1-4.

Images