Livestock breeding state monitoring method and system

Abstract

The invention provides a livestock breeding state monitoring method and system, and relates to the technical field of data processing, and the method comprises the steps: setting an RFID chip and a health monitoring sensor on each animal body; acquiring a multi-modal sensor data sequence of the animal in the time window; preprocessing the sensor data sequence; respectively extracting hidden states of various sensor data sequences through a long-short term memory neural network; based on a self-attention mechanism, adding a self-attention weight to the hidden state at each moment; obtaining intermediate monitoring results corresponding to the various sensor data sequences according to the hidden states and the self-attention weights of the various sensor data sequences; an independent channel is configured for each intermediate monitoring result in the full-connection network; based on a channel attention mechanism, adding a channel attention weight for each channel; and based on the weight of each channel, summarizing various intermediate monitoring results, and determining a comprehensive monitoring result.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method and system for monitoring animal husbandry status. Background Art

[0002] Livestock breeding status monitoring uses intelligent technology to achieve early disease warning, improve production efficiency, improve animal welfare and ensure food safety by tracking animal health, behavior and environmental conditions in real time. It helps the breeding model transform from extensive to precise and intelligent, reduces resource waste and environmental pollution, and promotes green breeding and sustainable development. At the same time, this technology provides a scientific decision-making basis for breeding management, enhances industry competitiveness, and supports the comprehensive upgrade of modern animal husbandry.

[0003] Currently, many livestock farms still rely on manual inspections to monitor livestock breeding status and make preliminary judgments by observing the animals' external behavior, food intake, body shape, mental state, etc. This is highly subjective, inefficient, and has poor real-time performance. It is difficult to meet the needs of large-scale breeding, especially when diseases are often discovered in the late stages.

[0004] With the rapid development of science and technology, more and more modern technologies are being applied to livestock breeding status monitoring, especially by installing cameras in breeding farms and then combining computer vision technology to monitor animal behavior (such as movement, feeding, abnormal behavior) to monitor livestock breeding status.

[0005] However, the accuracy of animal husbandry status monitoring technology using computer vision is affected by ambient light, occlusion, etc., and cannot directly reflect the physiological state, especially the early stages of the disease are difficult to accurately monitor through computer vision. At the same time, in a group breeding environment, a large number of animals are active at the same time, and the accuracy of individual segmentation and recognition is limited, especially for animals with similar appearance (such as pigs and cattle in the same batch). Identifying individuals may require the use of additional identification equipment (such as RFID, electronic ear tags), which increases the complexity of the system. Summary of the invention

[0006] In order to solve the technical problem that the existing livestock breeding status monitoring technology with the help of computer vision has the monitoring accuracy affected by ambient light, occlusion, etc., and cannot directly reflect the physiological state, especially the early stages of the disease are difficult to accurately monitor through computer vision, and at the same time, in a group breeding environment, a large number of animals are active at the same time, and the accuracy of individual segmentation and recognition is limited, especially for animals with similar appearance (such as pigs and cattle in the same batch), individual identification may require the use of additional identification equipment (such as RFID, electronic ear tags), which increases the complexity of the system, the present invention provides a livestock breeding status monitoring method and system.

[0007] The technical solution provided by the embodiment of the present invention is as follows:

[0008] First aspect:

[0009] An embodiment of the present invention provides a method for monitoring livestock breeding status, comprising:

[0010] S1: Install RFID chips and health monitoring sensors on each animal;

[0011] S2: Acquire the multimodal sensor data sequence of the animal in the time window;

[0012] S3: preprocessing the sensor data sequence;

[0013] S4: extracting the hidden states of various sensor data sequences respectively through a long short-term memory neural network;

[0014] S5: Based on the self-attention mechanism, add self-attention weights to the hidden states at each moment;

[0015] S6: performing state monitoring respectively according to the hidden states and self-attention weights of the various sensor data sequences to obtain intermediate monitoring results corresponding to the various sensor data sequences;

[0016] S7: configuring a separate channel for each intermediate monitoring result in the fully connected network, and obtaining hidden states of various sensor data sequences and corresponding intermediate monitoring results through the fully connected network;

[0017] S8: Based on the channel attention mechanism, add channel attention weights to each channel;

[0018] S9: Based on the weights of the various channels, the various intermediate monitoring results are summarized to determine a comprehensive monitoring result.

[0019] Second aspect:

[0020] An embodiment of the present invention provides a livestock breeding status monitoring system, comprising:

[0021] processor;

[0022] A memory having computer-readable instructions stored thereon, wherein when the computer-readable instructions are executed by the processor, the livestock breeding status monitoring method as described in the first aspect is implemented.

[0023] The third aspect:

[0024] An embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the livestock breeding status monitoring method as described in the first aspect is implemented.

[0025] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0026] In the present invention, the RFID chip provides a unique identity for each animal, ensuring the accuracy of individual monitoring and solving the problem of individual confusion in group farming. Based on the multimodal sensor data sequence, the key physiological data of the animal can be directly obtained, which can identify the early physiological signs of the disease and make up for the shortcomings of visual technology. Then, with the help of short-term memory neural network, the self-attention mechanism and the channel attention mechanism are introduced to achieve a comprehensive and integrated evaluation from behavior to physiology, avoid the error of a single information source, and improve the accuracy of livestock breeding status monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0028] Figure 1 A schematic diagram of a flow chart of a livestock breeding status monitoring method provided by an embodiment of the present invention;

[0029] Figure 2 A schematic diagram of the structure of a livestock breeding status monitoring system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0030] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0031] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "example" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of the word "example" is intended to present the concept in a specific way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either of the two.

[0032] In the embodiments of the present invention, "image" and "picture" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same. "of", "corresponding, relevant" and "corresponding" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same.

[0033] In order to make the technical problems, technical solutions and advantages to be solved by the present invention more clear, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0034] Reference Manual Attached Figure 1 , showing a flow chart of a livestock breeding status monitoring method provided by an embodiment of the present invention.

[0035] The embodiment of the present invention provides a method for monitoring the state of animal husbandry, which can be implemented by an animal husbandry status monitoring device, which can be a terminal or a server. The processing flow of the method for monitoring the state of animal husbandry can include the following steps:

[0036] S1: Install RFID chips and health monitoring sensors on each animal.

[0037] Among them, RFID (Radio Frequency Identification) chip is a technology that transmits data through radio waves for identification and communication. In animal husbandry, RFID chips are often used for animal identification and information tracking. It is a non-contact automatic recognition technology that can quickly and accurately obtain information about the target object.

[0038] Among them, health monitoring sensors include: position sensors, acoustic sensors, temperature sensors, heart rate sensors and food intake sensors.

[0039] Among them, the position sensor is used to monitor the animal's position, activity trajectory and behavior pattern. Optionally, the RFID chip itself can also be used as a position sensor.

[0040] Among them, acoustic sensors are used to analyze the health status and behavior of animals by capturing the sounds of animals or changes in the surrounding sound field.

[0041] Among them, temperature sensors are used to measure the body temperature or ambient temperature of animals and assess health and comfort. Built-in temperature sensors (such as thermistors or infrared sensors) detect the temperature on the surface or inside the body of animals. This is usually achieved through ear tags, subcutaneous implants or collars.

[0042] Among them, heart rate sensors are used to monitor the heart rate of animals in real time and assess their health and emotional state. The light source and photodetector detect changes in light absorption caused by blood flow and calculate the heart rate. They are often used in ear tags, collars or skin patch devices.

[0043] Among them, feeding sensors monitor the animal's feeding behavior, food intake, and water intake to optimize feeding management. Specifically, vibration sensors or acoustic sensors can be installed on collars or ear tags to detect chewing frequency and duration. Weighing sensors can also be installed in the trough to measure the amount of feed reduction per feeding.

[0044] S2: Obtain a multimodal sensor data sequence of the animal under a time window.

[0045] Correspondingly, the sensor data sequence includes: an exercise data sequence, a call data sequence, a body temperature data sequence, a heart rate data sequence and a food intake data sequence.

[0046] Among them, those skilled in the art can set the width of the time window according to actual conditions, and the present invention does not limit it.

[0047] S3: Preprocess the sensor data sequence.

[0048] Optionally, preprocessing includes: noise denoising, outlier removal, missing value filling, time series alignment and normalization.

[0049] Among them, noise denoising is to remove meaningless random interference or error signals from sensor data and retain key valid information. It can be implemented by using filtering algorithms.

[0050] Among them, outlier removal is to detect and remove abnormal points that do not conform to data characteristics or distribution. Specifically, the Z-score method can be used.

[0051] Missing value filling is to use a reasonable method to fill in the missing data caused by sensor failure or signal interruption. Specifically, the previous valid value can be used to fill the missing point, which is suitable for steadily changing data.

[0052] Among them, timing alignment is to align the data of different sensors to a unified time axis to ensure that they can be analyzed within the same time window.

[0053] Among them, normalization converts the data to a uniform scale or range (such as [0,1] or [-1,1]) to eliminate the differences between different data dimensions or dimensions.

[0054] S4: Through the long short-term memory neural network, the hidden states of various sensor data sequences are extracted respectively.

[0055] Among them, the Long Short-Term Memory (LSTM) neural network is a recurrent neural network (RNN) variant specifically designed to process and model time series data. LSTM solves the gradient vanishing and gradient exploding problems faced by traditional RNN in modeling long-term dependencies by introducing gating mechanisms and cell states.

[0056] In a possible implementation, S4 specifically includes: extracting the hidden states of various sensor data sequences respectively through a long short-term memory neural network according to the following formula:

[0057] I t =Sigmoid(W XI X t +W HI h t-1 +b I )

[0058] F t =Sigmoid(W XF X t +W HF h t-1 +b F )

[0059] O t =Sigmoid(W XO X t +W HO h t-1 +b O )

[0060]

[0061] h t =O t tanh(C t )

[0062] Among them, X t represents the sensor data at time t in the sensor data sequence, I t represents the activation output vector of the input gate at time t, Sigmoid() represents the Sigmoid activation function, and W XI Represents the weight matrix between the word segmentation sequence and the input gate, W HI represents the weight matrix between the hidden state and the input gate, b I represents the bias term of the input gate, F t represents the activation output vector of the forget gate at time t, W XF Represents the weight matrix between the word segmentation sequence and the forget gate, W HF represents the weight matrix between the hidden state and the forget gate, b F represents the bias term of the forget gate, O t represents the activation output vector of the output gate at time t, W XO Represents the weight matrix between the word segmentation sequence and the output gate, W HO represents the weight matrix between the hidden state and the output gate, b O represents the bias term of the output gate, C t represents the activation output vector of the cell storage unit at time t, represents the temporary cell state of the cell storage unit at time t, C t-1represents the activation output vector of the cell storage unit at time t-1, tanh() represents the tanh activation function, W XC Represents the weight matrix between the word segmentation sequence and the cell storage unit, W HC represents the weight matrix between the hidden state and the cell storage unit, b C represents the bias term of the cell storage unit, h t represents the hidden state at time t, h t-1 represents the hidden state at time t-1, and t represents time.

[0063] In the present invention, by using LSTM to extract the hidden state of the sensor data sequence, its advantages in time series modeling can be fully utilized, especially when processing complex data with multimodal, long-term dependencies and nonlinear characteristics. LSTM provides an accurate and efficient feature extraction method, laying a solid foundation for subsequent state monitoring and decision-making.

[0064] S5: Based on the self-attention mechanism, add self-attention weights to the hidden states at each moment.

[0065] In a possible implementation, S5 is specifically: based on the self-attention mechanism, according to the following formula, a self-attention weight is added to the hidden state at each moment:

[0066] e t =σ(W h h t +b h )

[0067]

[0068] Among them, e t represents the attention score at time t, σ() represents the activation function, and W h represents the self-attention weight matrix, h t represents the hidden state at time t, b h represents the self-attention bias term, α t represents the self-attention weight at time t, exp() represents an exponential function with a natural constant as the base, and T represents the total duration in the time window.

[0069] In the present invention, the attention weight calculated by the self-attention mechanism assigns a dynamic weight to each moment in the time series, reflecting the importance of the moment in the overall time series. The hidden state at some moments may contain more critical information for the final task (such as health monitoring or anomaly detection), while the hidden state at other moments may be noisy. The attention weight can highlight the information of these critical moments and weaken the influence of unimportant moments.

[0070] S6: According to the hidden states and self-attention weights of various sensor data sequences, state monitoring is performed respectively to obtain intermediate monitoring results corresponding to various sensor data sequences.

[0071] In a possible implementation, S6 specifically includes: obtaining an intermediate monitoring result according to the following formula:

[0072]

[0073] Among them, P t represents the intermediate monitoring result at time t, Softmax() represents the Softmax activation function, W p represents the classification weight matrix, α t represents the self-attention weight at time t, h t represents the hidden state at time t, b p Represents the classification bias term, and T represents the total duration in the time window.

[0074] In the present invention, the Softmax function normalizes the weighted summation result into a probability distribution, so that the intermediate monitoring result can clearly represent the probability of belonging to different states (such as normal, abnormal or specific behavior). In the above way, the key features of the time series can be efficiently extracted to generate accurate and multi-category supported intermediate monitoring results. This method not only improves the accuracy, flexibility and interpretability of state monitoring, but also provides high-quality input for subsequent multimodal fusion and decision-making.

[0075] S7: Configure a separate channel for each intermediate monitoring result in the fully connected network, and obtain the hidden states of various sensor data sequences and the corresponding intermediate monitoring results through the fully connected network.

[0076] S8: Based on the channel attention mechanism, add channel attention weights to each channel.

[0077] In a possible implementation, S8 specifically includes sub-steps S801 and S802:

[0078] S801: compress the hidden state of the sensor data sequence extracted in each channel:

[0079]

[0080] Among them, y represents the compressed feature, y(k) represents the compressed feature of the kth channel, and h t represents the hidden state at time t, h t (k) represents the hidden state of the kth channel at time t, and T represents the total duration of the time window.

[0081] S802: Reconstruct the compressed features and add channel attention weights to each channel:

[0082] ω k =Sigmoid{W s [ReLU(W r (y(k)))]}

[0083] Among them, ω k represents the channel attention weight of the kth channel, Sigmoid represents the Sigmoid activation function, ReLU represents the ReLU activation function, and W s represents the reconstruction of the fully connected layer, W r represents a compressed fully connected layer.

[0084] In the present invention, global information is extracted by compressing the time series features of each channel, and then the importance of key channels is dynamically highlighted through nonlinear reconstruction and allocation of channel attention weights, while the influence of irrelevant or noisy channels is suppressed. This design not only reduces the computational complexity, but also effectively enhances the fusion effect of multimodal data, improves the adaptability and robustness of the model to multiple sensor data, and provides better interpretability through the distribution of attention weights, making it easier to understand the contribution of each sensor in the monitoring task.

[0085] S9: Based on the weights of each channel, various intermediate monitoring results are summarized to determine the comprehensive monitoring results.

[0086] In a possible implementation, S9 specifically includes: summarizing various intermediate monitoring results based on the weights of the various channels, and determining a comprehensive monitoring result according to the following formula:

[0087]

[0088] Among them, F t represents the comprehensive monitoring result at time t, P t represents the intermediate monitoring result at time t, P t (k) represents the intermediate monitoring result of the kth channel at time t, ω k represents the channel attention weight of the kth channel, and K represents the total number of channels.

[0089] Furthermore, the category with the largest probability value in the probability distribution represented by the comprehensive monitoring result can be taken as the comprehensive monitoring result.

[0090] In the present invention, through the weighted aggregation of channel attention weights and intermediate monitoring results, the model can efficiently fuse multimodal sensor data, dynamically highlight key channels, and weaken the impact of noise, thereby generating more accurate, comprehensive, robust and interpretable comprehensive monitoring results, and thus timely discover health problems of farmed animals. This method is suitable for complex and diverse livestock monitoring scenarios, and is both real-time and flexible.

[0091] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0092] In the present invention, the RFID chip provides a unique identity for each animal, ensuring the accuracy of individual monitoring and solving the problem of individual confusion in group farming. Based on the multimodal sensor data sequence, the key physiological data of the animal can be directly obtained, which can identify the early physiological signs of the disease and make up for the shortcomings of visual technology. Then, with the help of short-term memory neural network, the self-attention mechanism and the channel attention mechanism are introduced to achieve a comprehensive and integrated evaluation from behavior to physiology, avoid the error of a single information source, and improve the accuracy of livestock breeding status monitoring.

[0093] Reference Manual Attached Figure 2 , showing a structural schematic diagram of a livestock breeding status monitoring system provided by the present invention.

[0094] The present invention also provides a livestock breeding state monitoring system 20, which is applied to the above-mentioned livestock breeding state monitoring method, comprising:

[0095] Processor 201.

[0096] The memory 202 stores computer-readable instructions. When the computer-readable instructions are executed by the processor 201, the livestock breeding status monitoring method of the method embodiment is implemented.

[0097] The livestock breeding status monitoring system 20 provided by the present invention can execute the above-mentioned livestock breeding status monitoring method and achieve the same or similar technical effects. To avoid repetition, the present invention will not go into details.

[0098] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0099] In the present invention, the RFID chip provides a unique identity for each animal, ensuring the accuracy of individual monitoring and solving the problem of individual confusion in group farming. Based on the multimodal sensor data sequence, the key physiological data of the animal can be directly obtained, which can identify the early physiological signs of the disease and make up for the shortcomings of visual technology. Then, with the help of short-term memory neural network, the self-attention mechanism and the channel attention mechanism are introduced to achieve a comprehensive and integrated evaluation from behavior to physiology, avoid the error of a single information source, and improve the accuracy of livestock breeding status monitoring.

[0100] It should be understood that the processor in the embodiment of the present invention may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0101] It should also be understood that the memory in the embodiments of the present invention may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic random access memory (DRAM), synchronous DRAM (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link DRAM (SLDRAM), and direct RAM bus random access memory (DR RAM).

[0102] The above embodiments can be implemented in whole or in part by software, hardware (such as circuits), firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a tape), an optical medium (for example, a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0103] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship, but it may also indicate an "and / or" relationship. Please refer to the context for specific understanding.

[0104] In the present invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can be represented by: a, b, c, ab, ac, bc, or abc, where a, b, c can be single or multiple.

[0105] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0106] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0107] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described equipment, devices and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0108] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0109] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0110] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0111] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage media include: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, etc., various media that can store program codes.

[0112] An embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, the livestock breeding status monitoring method as described in the method embodiment is implemented.

[0113] A computer-readable storage medium provided by the present invention can implement the steps and effects of the livestock breeding status monitoring method of the above method embodiment. To avoid repetition, the present invention will not go into details.

[0114] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0115] In the present invention, the RFID chip provides a unique identity for each animal, ensuring the accuracy of individual monitoring and solving the problem of individual confusion in group farming. Based on the multimodal sensor data sequence, the key physiological data of the animal can be directly obtained, which can identify the early physiological signs of the disease and make up for the shortcomings of visual technology. Then, with the help of short-term memory neural network, the self-attention mechanism and the channel attention mechanism are introduced to achieve a comprehensive and integrated evaluation from behavior to physiology, avoid the error of a single information source, and improve the accuracy of livestock breeding status monitoring.

[0116] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art who is familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

[0117] There are a few points to note:

[0118] (1) The drawings of the embodiments of the present invention only relate to the structures related to the embodiments of the present invention, and other structures may refer to the general design.

[0119] (2) For the sake of clarity, in the drawings used to describe the embodiments of the present invention, the thickness of the layers or regions is exaggerated or reduced, that is, these drawings are not drawn according to the actual scale. It is understood that when an element such as a layer, film, region or substrate is referred to as being "on" or "under" another element, the element may be "directly" "on" or "under" the other element or there may be intermediate elements.

[0120] (3) In the absence of conflict, the embodiments of the present invention and the features therein may be combined with each other to obtain new embodiments.

[0121] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. The protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for monitoring livestock breeding status, characterized in that: include: S1: Install RFID chips and health monitoring sensors on each animal; S2: Acquire the multimodal sensor data sequence of the animal in the time window; S3: preprocessing the sensor data sequence; S4: extracting the hidden states of various sensor data sequences respectively through a long short-term memory neural network; S5: Based on the self-attention mechanism, add self-attention weights to the hidden states at each moment; S6: performing state monitoring respectively according to the hidden states and self-attention weights of the various sensor data sequences to obtain intermediate monitoring results corresponding to the various sensor data sequences; S7: configuring a separate channel for each intermediate monitoring result in the fully connected network, and obtaining hidden states of various sensor data sequences and corresponding intermediate monitoring results through the fully connected network; S8: Based on the channel attention mechanism, add channel attention weights to each channel; S9: Based on the weights of the various channels, the various intermediate monitoring results are summarized to determine a comprehensive monitoring result.

2. The livestock breeding status monitoring method according to claim 1, characterized in that: The health monitoring sensors include: a position sensor, an acoustic sensor, a temperature sensor, a heart rate sensor, and a food intake sensor; Correspondingly, the sensor data sequence includes: an exercise data sequence, a call data sequence, a body temperature data sequence, a heart rate data sequence and a food intake data sequence.

3. The livestock breeding status monitoring method according to claim 1, characterized in that: The preprocessing includes: noise denoising, outlier removal, missing value filling, time series alignment and normalization processing.

4. The livestock breeding status monitoring method according to claim 1, characterized in that: The S4 is specifically: Through the long short-term memory neural network, the hidden states of various sensor data sequences are extracted according to the following formula: I t =Sigmoid(W XI X t +W HI h t-1 +b I ) F t =Sigmoid(W XF X t +W HF h t-1 +b F ) O t =Sigmoid(W XO X t +W HO h t-1 +b O ) h t =O t ·tanh(C t ) Among them, X t represents the sensor data at time t in the sensor data sequence, I t represents the activation output vector of the input gate at time t, Sigmoid() represents the Sigmoid activation function, and W XI Represents the weight matrix between the word segmentation sequence and the input gate, W HI represents the weight matrix between the hidden state and the input gate, b I represents the bias term of the input gate, F t represents the activation output vector of the forget gate at time t, W XF Represents the weight matrix between the word segmentation sequence and the forget gate, W HF represents the weight matrix between the hidden state and the forget gate, b F represents the bias term of the forget gate, O t represents the activation output vector of the output gate at time t, W XO Represents the weight matrix between the word segmentation sequence and the output gate, W HO represents the weight matrix between the hidden state and the output gate, b O represents the bias term of the output gate, C t represents the activation output vector of the cell storage unit at time t, represents the temporary cell state of the cell storage unit at time t, C t-1 represents the activation output vector of the cell storage unit at time t-1, tanh() represents the tanh activation function, W XC Represents the weight matrix between the word segmentation sequence and the cell storage unit, W HC represents the weight matrix between the hidden state and the cell storage unit, b C represents the bias term of the cell storage unit, h t represents the hidden state at time t, h t-1 represents the hidden state at time t-1, and t represents time.

5. The livestock breeding status monitoring method according to claim 4, characterized in that: The S5 is specifically: Based on the self-attention mechanism, the self-attention weight is added to the hidden state at each moment according to the following formula: e t =σ(W h h t +b h ) Among them, e t represents the attention score at time t, σ() represents the activation function, and W h represents the self-attention weight matrix, h t represents the hidden state at time t, b h represents the self-attention bias term, α t represents the self-attention weight at time t, exp() represents an exponential function with a natural constant as the base, and T represents the total duration in the time window.

6. The livestock breeding status monitoring method according to claim 4, characterized in that: The S6 is specifically: According to the following formula, the intermediate monitoring result is obtained: Among them, P t represents the intermediate monitoring result at time t, Softmax() represents the Softmax activation function, W p represents the classification weight matrix, α t represents the self-attention weight at time t, h t represents the hidden state at time t, b p Represents the classification bias term, and T represents the total duration in the time window.

7. The livestock breeding status monitoring method according to claim 6, characterized in that: The S8 specifically includes: S801: compress the hidden state of the sensor data sequence extracted in each channel: Among them, y represents the compressed feature, y(k) represents the compressed feature of the kth channel, and h t represents the hidden state at time t, h t (k) represents the hidden state of the kth channel at time t, and T represents the total duration in the time window; S802: Reconstruct the compressed features and add channel attention weights to each channel: ω k =Sigmoid{W s [ReLU(W r (y(k)))]} Among them, ω k represents the channel attention weight of the kth channel, Sigmoid represents the Sigmoid activation function, ReLU represents the ReLU activation function, and W s represents the reconstruction of the fully connected layer, W r represents a compressed fully connected layer.

8. The livestock breeding status monitoring method according to claim 7, characterized in that: The S9 specifically includes: Based on the weight of each channel, the various intermediate monitoring results are summarized and the comprehensive monitoring result is determined according to the following formula: Among them, F t represents the comprehensive monitoring result at time t, P t represents the intermediate monitoring result at time t, P t (k) represents the intermediate monitoring result of the kth channel at time t, ω k represents the channel attention weight of the kth channel, and K represents the total number of channels.

9. A livestock breeding status monitoring system, characterized in that: include: processor; A memory having computer-readable instructions stored thereon, wherein when the computer-readable instructions are executed by the processor, the livestock breeding status monitoring method as described in any one of claims 1 to 8 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the livestock breeding status monitoring method as described in any one of claims 1 to 8 is implemented.

Images