A method and system for monitoring livestock production environment based on the Internet of Things

Abstract

This invention belongs to the field of environmental monitoring technology and discloses a method and system for monitoring livestock production environment based on the Internet of Things (IoT). The method includes: deploying sensor modules in different areas of the farm to collect data on temperature, humidity, light intensity, and harmful gas concentrations, and performing data preprocessing; using edge computing technology to perform preliminary processing on the preprocessed data, extracting key features, and transmitting the data to a cloud platform; establishing a regionally differentiated monitoring model on the cloud platform to monitor each key feature data in real time; triggering an early warning mechanism if abnormal fluctuations in temperature, humidity, or harmful gas concentrations are detected, notifying management personnel for timely handling; and deploying a data fusion algorithm on the cloud platform to integrate multi-region sensor data and generate a global environmental parameter distribution map. This invention effectively solves the challenges of data collection, transmission, and analysis in complex environments of livestock farms, achieving refined environmental management and improving breeding efficiency and safety.

Description

Technical Field

[0001] This invention belongs to the field of environmental monitoring technology, specifically relating to a method and system for monitoring the livestock production environment based on the Internet of Things. Background Technology

[0002] In livestock production environment monitoring, deployed sensors need to collect key environmental parameters such as temperature, humidity, light intensity, and harmful gas concentrations in real time. These sensors are typically installed in different areas of the farm to ensure the comprehensiveness and representativeness of the data. However, due to the complex environment of farms, sensors face several technical challenges in actual operation. First, sensors need to operate stably and continuously under conditions of high humidity, high dust, and extreme temperatures, placing extremely high demands on their durability and accuracy. Second, the data collected by the sensors needs to be transmitted to the management cloud platform via a wireless network, but within the farm, wireless signals may be interfered with by buildings, equipment, and animal activities, leading to unstable or delayed data transmission. Furthermore, sensors may experience insufficient battery power or equipment aging after prolonged operation, affecting the continuity and reliability of the data. Simultaneously, environmental parameters may vary significantly between different areas of the farm; for example, poorly ventilated areas may accumulate harmful gases, while well-lit areas may experience excessively high temperatures. This necessitates sensors capable of differentiated monitoring for different areas, but existing systems often struggle to achieve this level of refined management. Finally, the management cloud platform needs to process and analyze massive amounts of sensor data in real time. However, given the sheer volume and diverse sources of this data, efficiently integrating and processing it while avoiding redundancy and errors remains a critical technical challenge. These issues collectively constitute the core challenges in livestock production environment monitoring, requiring gradual solutions through technological and system optimization. Summary of the Invention

[0003] To address the problems existing in the prior art, this invention provides a method and system for monitoring the livestock production environment based on the Internet of Things (IoT). This effectively solves the challenges of data collection, transmission, and analysis in complex environments of farms, enabling refined environmental management and improving breeding efficiency and safety.

[0004] To achieve the above objectives, the present invention provides the following solution:

[0005] A method for monitoring the livestock production environment based on the Internet of Things, the method comprising:

[0006] Sensor modules were deployed in different areas of the farm to collect data on temperature, humidity, light intensity, and concentration of harmful gases, and the data was preprocessed.

[0007] Edge computing technology is used to perform preliminary processing on the pre-processed data, extract key features, and then transmit the data to the cloud platform.

[0008] Establish a regionally differentiated monitoring model in the cloud platform to monitor key characteristic data in real time. If abnormal fluctuations in temperature, humidity or concentration of harmful gases are detected, trigger the early warning mechanism and notify the management personnel to handle the situation in a timely manner.

[0009] Deploy data fusion algorithms in the cloud platform to integrate sensor data from multiple regions and generate a global environmental parameter distribution map.

[0010] Preferably, the method for deploying sensor modules in different areas of the farm includes:

[0011] The preprocessed data is transmitted through low-power wide-area network technology. Specifically, relay nodes are deployed in the farm, and the location and number of nodes are planned in a reasonable manner.

[0012] Establish a data caching mechanism in the cloud platform. If a data packet is detected to be lost or delayed, the missing data will be automatically resent from the cache.

[0013] Based on cached data, multiple types of sensors are deployed to address environmental differences in different regions.

[0014] Preferably, edge computing technology is used to perform preliminary processing on the preprocessed data and extract key features. The method includes: extracting key features by constructing an edge computing network architecture framework, wherein the edge computing network architecture framework includes: service layer, platform layer, transmission layer, edge layer, and perception layer.

[0015] The perception layer, connected to the edge gateway, can transmit various environmental information to the edge layer in real time.

[0016] The edge layer mainly consists of edge gateways, edge cloud, and edge controllers. The edge gateway is responsible for managing various sensors and computing resources. The edge cloud provides storage and computing resources for monitoring data, enabling data processing and analysis to be performed locally. The edge controller is responsible for real-time control and optimization of various sensors.

[0017] The transport layer is primarily responsible for data transmission, selecting the appropriate network for information transmission based on different application scenarios.

[0018] The platform layer serves as the hub for data collection and processing, analyzing various types of environmental data;

[0019] The service layer focuses on the analysis of environmental monitoring data. Through in-depth analysis of environmental monitoring data, the service layer can promptly identify abnormal information. The service layer also predicts environmental trends in the future by mining and analyzing historical data.

[0020] Discrete wavelet transform is introduced to achieve time-frequency domain feature extraction.

[0021] X(a)=∑(W j，m ·Ψ(a-mj))

[0022] In the formula, X(a) represents the original data, and W j,m Let Ψ(a-mj) represent the wavelet coefficients, Ψ(a-mj) represent the wavelet basis functions, j represent the scale, and m represent the translation.

[0023] Preferred methods for establishing regionally differentiated monitoring models in a cloud platform include:

[0024] An edge computing network architecture is deployed at sensor nodes to acquire environmental data, preprocess the data, and filter out noise.

[0025] Feature extraction algorithms are used to extract key features from preprocessed data to obtain a feature dataset, which is then transmitted to the cloud platform.

[0026] The system receives feature data on the cloud platform and uses a data integrity detection algorithm to determine if data is lost during transmission. If data loss is detected, a retransmission mechanism is triggered.

[0027] Based on the feature dataset, a clustering algorithm is used to classify the environmental data and determine the distribution patterns of environmental features in different regions.

[0028] Based on the distribution patterns of environmental characteristics, a regional division algorithm is adopted to generate a regionally differentiated environmental monitoring model.

[0029] Preferably, methods for establishing regionally differentiated monitoring models in a cloud platform to monitor key characteristic data in real time include:

[0030] Establish a conceptual model: In the formula, R, f(H), f(E), f(R), and f(V) represent the regional environmental anomaly, the risk of the anomaly source, the receptor exposure, the receptor resistance, and the receptor vulnerability of the evaluation object, respectively.

[0031] Each evaluation indicator is dimensionless: x ij (t k ) for object s i The j-th indicator in t k Given the values ​​at time points (i = 1, 2, ..., n; j = 1, 2, ..., m; k = 1, 2, ..., N), we have: In the formula, x′ ij (t k () represents the dimensionless value; and σ j (t k ) represent the j-th indicator at t k Mean and standard deviation at time points;

[0032] The method of grading by both horizontal and vertical separation is used to calculate and obtain the t of each evaluation object. k The risk index of abnormal sources, receptor exposure, and resilience index at any given time;

[0033] If abnormal fluctuations in temperature, humidity, or harmful gas concentration are detected, an early warning mechanism will be triggered, and management personnel will be notified to handle the situation promptly.

[0034] Preferably, methods for deploying data fusion algorithms in a cloud platform to integrate multi-regional sensor data and generate a global environmental parameter distribution map include:

[0035] For each region, the K-nearest neighbor classification algorithm is used to match the dimensionality-reduced sensor data; data that closely matches the preset samples in the storage is selected, and then the filtered data is precisely matched by a triplet + neural network with ResNet as the backbone network, thereby obtaining the fused sensor data for each region;

[0036] All sensor data from each region after fusion are transmitted to the central fusion module, where tracking, correlation, and estimation are performed to generate a global environmental parameter distribution map. The central fusion module uses the FOP-MOC model.

[0037] The present invention also provides an Internet of Things-based livestock production environment monitoring system, the system being used to implement any of the methods described above, the system comprising: a preprocessing module, a feature extraction module, an anomaly detection module, and a data fusion module;

[0038] The preprocessing module is used to deploy sensor modules in different areas of the farm to collect data on temperature, humidity, light intensity and concentration of harmful gases and perform data preprocessing.

[0039] The feature extraction module is used to perform preliminary processing on the preprocessed data using edge computing technology, extract key features, and then transmit them to the cloud platform.

[0040] The anomaly detection module is used to establish a regionally differentiated monitoring model in the cloud platform, monitor key characteristic data in real time, and trigger an early warning mechanism if abnormal fluctuations in temperature, humidity or harmful gas concentration are detected, so as to notify the management personnel to handle the situation in a timely manner.

[0041] The data fusion module is used to deploy data fusion algorithms in the cloud platform, integrate sensor data from multiple regions, and generate a global environmental parameter distribution map.

[0042] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0043] This invention discloses a method and system for monitoring the livestock production environment based on the Internet of Things (IoT). Addressing the characteristics of livestock farms—high humidity, high dust levels, and large temperature fluctuations—it employs sensor modules resistant to extreme environments to collect environmental data in different areas. Data preprocessing is performed using an integrated adaptive filtering algorithm, and transmission is optimized using low-power wide-area network (LPWAN) technology and relay nodes. This invention establishes a data caching and retransmission mechanism on a cloud platform to ensure data continuity. Combining edge computing technology and a regionally differentiated monitoring model, real-time anomaly monitoring and early warning are achieved. Simultaneously, a global environmental parameter distribution map is provided through equipment health monitoring and data fusion algorithms. This invention effectively solves the challenges of data collection, transmission, and analysis in complex livestock farm environments, enabling refined environmental management and improving livestock farming efficiency and safety. Attached Figure Description

[0044] To more clearly illustrate the technical solution of the present invention, the drawings used in the embodiments are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1 This is a schematic diagram of a method for monitoring the livestock production environment based on the Internet of Things, according to an embodiment of the present invention.

[0046] Figure 2 This is a schematic diagram of the edge computing network architecture framework according to an embodiment of the present invention. Detailed Implementation

[0047] 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. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0048] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0049] Example 1

[0050] like Figure 1 As shown, this invention provides a method for monitoring the livestock production environment based on the Internet of Things (IoT), the method comprising:

[0051] Sensor modules were deployed in different areas of the farm to collect data on temperature, humidity, light intensity, and concentration of harmful gases, and the data was preprocessed.

[0052] Edge computing technology is used to perform preliminary processing on the pre-processed data, extract key features, and then transmit the data to the cloud platform.

[0053] Establish a regionally differentiated monitoring model in the cloud platform to monitor key characteristic data in real time. If abnormal fluctuations in temperature, humidity or concentration of harmful gases are detected, trigger the early warning mechanism and notify the management personnel to handle the situation in a timely manner.

[0054] Deploy data fusion algorithms in the cloud platform to integrate sensor data from multiple regions and generate a global environmental parameter distribution map.

[0055] In this embodiment, an adaptive filtering algorithm is integrated into the sensor module to preprocess data in high-humidity and high-dust environments, remove noise interference, and improve data accuracy, including:

[0056] The sensor module is deployed in the high-humidity and high-dust area of ​​the farm to collect temperature, humidity, light and gas data;

[0057] An adaptive filtering algorithm is used to preprocess the original data to remove noise interference, resulting in a filtered dataset.

[0058] Based on a preset threshold, determine whether the filtered temperature value exceeds the threshold.

[0059] If the temperature value exceeds the threshold, a temperature anomaly warning is triggered, and the filtered humidity value and gas value are obtained for correlation analysis.

[0060] If neither the humidity nor the gas value exceeds the preset threshold, only the temperature anomaly information will be recorded and the data set will be updated.

[0061] If the humidity or gas value exceeds the preset threshold, a comprehensive early warning will be triggered, and a comprehensive anomaly report will be generated.

[0062] The K-means clustering algorithm was used to classify the comprehensive anomaly reports, determine the anomaly types and frequencies, and generate environmental monitoring results for the aquaculture farm.

[0063] In this embodiment, the method for deploying sensor modules in different areas of the farm includes:

[0064] The preprocessed data is transmitted through low-power wide-area network technology. Specifically, relay nodes are deployed in the farm, and the location and number of nodes are planned in a reasonable manner.

[0065] Establish a data caching mechanism in the cloud platform. If a data packet is detected to be lost or delayed, the missing data will be automatically resent from the cache.

[0066] Based on cached data, multiple types of sensors are deployed to address environmental differences in different regions.

[0067] Specifically, the arrangement of relay nodes within the farm, and the rational planning of node locations and numbers, include:

[0068] Based on the location and number of nodes, the signal coverage and strength are calculated to determine if there are any signal blind spots. If blind spots exist, the node locations and numbers are adjusted. Signal coverage and strength data are acquired, and the relay node layout is optimized in conjunction with the distribution of buildings and equipment to reduce signal interference. Filtered data is transmitted through the optimized relay node layout, and signal transmission quality is monitored to determine if there is any data loss or delay. If data loss or delay is detected, the relay node layout is readjusted, and signal transmission quality is monitored again until the preset transmission standards are met. Clustering algorithms are used to classify the transmission quality data, determine the distribution patterns of signal coverage and interference, and generate signal coverage and interference distribution maps. Based on the signal coverage and interference distribution maps, the relay node layout is updated to continuously optimize data transmission efficiency and signal coverage.

[0069] Specifically, based on the data packet content in the buffer, multiple types of sensors are deployed in the transmission chain to acquire environmental data for each area. For the acquired environmental data, detection points monitor the integrity and timeliness of the data packets, recording the amount of data loss and latency. If a detection point detects data packet loss or latency, a retransmitter extracts the missing values ​​from the buffer and resends them to the transmission chain. Based on the retransmitter's operational data, the environmental monitoring status of each area is updated, clarifying regionally differentiated environmental standards. Clustering algorithms are used to classify the amount of data loss and latency, obtaining the data anomaly distribution patterns for different areas. Based on these data anomaly distribution patterns, the buffer's storage strategy is adjusted, and the regionally differentiated monitoring model is optimized. Through the optimized monitoring model, the sensor deployment strategy is reconfigured to achieve refined environmental management.

[0070] In this embodiment, as Figure 2 As shown, edge computing technology is used to perform preliminary processing on the preprocessed data and extract key features. The methods include: extracting key features by constructing an edge computing network architecture framework, which includes: service layer, platform layer, transmission layer, edge layer, and perception layer.

[0071] The perception layer, connected to the edge gateway, can transmit various environmental information to the edge layer in real time.

[0072] The edge layer mainly consists of edge gateways, edge cloud, and edge controllers. The edge gateway is responsible for managing various sensors and computing resources. The edge cloud provides storage and computing resources for monitoring data, enabling data processing and analysis to be performed locally. The edge controller is responsible for real-time control and optimization of various sensors.

[0073] The transport layer is primarily responsible for data transmission, selecting the appropriate network for information transmission based on different application scenarios.

[0074] The platform layer serves as the hub for data collection and processing, analyzing various types of environmental data;

[0075] The service layer focuses on the analysis of environmental monitoring data. Through in-depth analysis of environmental monitoring data, the service layer can promptly identify abnormal information. The service layer also predicts environmental trends in the future by mining and analyzing historical data.

[0076] Discrete wavelet transform is introduced to achieve time-frequency domain feature extraction.

[0077] X(a)=∑(W j，m ·Ψ(a-mj))

[0078] In the formula, X(a) represents the original data, and W j，m Let Ψ(a-mj) represent the wavelet coefficients, Ψ(a-mj) represent the wavelet basis functions, j represent the scale, and m represent the translation.

[0079] In this embodiment, the method for establishing a regionally differentiated monitoring model in a cloud platform includes:

[0080] An edge computing network architecture is deployed at sensor nodes to acquire environmental data, preprocess the data, and filter out noise.

[0081] Feature extraction algorithms are used to extract key features from preprocessed data to obtain a feature dataset, which is then transmitted to the cloud platform.

[0082] The system receives feature data on the cloud platform and uses a data integrity detection algorithm to determine if data is lost during transmission. If data loss is detected, a retransmission mechanism is triggered.

[0083] Based on the feature dataset, a clustering algorithm is used to classify the environmental data and determine the distribution patterns of environmental features in different regions.

[0084] Based on the distribution patterns of environmental characteristics, a regional division algorithm is adopted to generate a regionally differentiated environmental monitoring model.

[0085] In this embodiment, the method for establishing a regionally differentiated monitoring model in the cloud platform and monitoring key feature data in real time includes:

[0086] Establish a conceptual model: In the formula, R, f(H), f(E), f(R), and f(V) represent the regional environmental anomaly, the risk of the anomaly source, the receptor exposure, the receptor resistance, and the receptor vulnerability of the evaluation object, respectively.

[0087] Each evaluation indicator is dimensionless: x ij (t k ) for object s i The j-th indicator in t k Given the values ​​at time points (i = 1, 2, ..., n; j = 1, 2, ..., m; k = 1, 2, ..., N), we have: In the formula, x′ ij (t k () represents the dimensionless value; and σ j (t k ) represent the j-th indicator at t k Mean and standard deviation at time points;

[0088] The method of grading by both horizontal and vertical separation is used to calculate and obtain the t of each evaluation object. k The risk index of abnormal sources, receptor exposure, and resilience index at any given time;

[0089] If abnormal fluctuations in temperature, humidity, or harmful gas concentration are detected, an early warning mechanism will be triggered, and management personnel will be notified to handle the situation promptly.

[0090] Specifically, the method of grading by both horizontal and vertical dimensions is used to calculate and obtain the t-values ​​of each evaluation object. k Methods for determining the risk index of abnormal sources, receptor exposure, and resilience index at any given time include:

[0091] (1) For a given time t k n evaluation objects and m indicators x1, x2, ..., x m The numerical value (dimensionless) is represented by matrix A. k This means, that is:

[0092]

[0093] (2) Calculate the m×m real symmetric matrix H k H k =A k T A k ;

[0094] (3) Calculate H k The largest eigenvalue λ max The corresponding weight coefficient vectors are obtained and normalized to calculate the weight ω of each indicator. j (t k );

[0095] (4) Calculation function:

[0096]

[0097] In the formula: yi (t k ) is the evaluation object s i In t k The abnormal source risk index, receptor exposure index, and receptor resistance index at any given time.

[0098] (5) Calculate the environmental anomaly index. Based on the regional environmental anomaly conceptual model established in this paper, calculate the evaluation object s. i In t k Time-region environmental anomaly index R i (t k ) and receptor vulnerability index V i (t k ).

[0099] In this embodiment, the method for deploying a data fusion algorithm in a cloud platform to integrate multi-region sensor data and generate a global environmental parameter distribution map includes:

[0100] For each region, the K-nearest neighbor classification algorithm is used to match the dimensionality-reduced sensor data; data that closely matches the preset samples in the storage is selected, and then the filtered data is precisely matched by a triplet + neural network with ResNet as the backbone network, thereby obtaining the fused sensor data for each region;

[0101] All sensor data from each region after fusion are transmitted to the central fusion module, where tracking, correlation, and estimation are performed to generate a global environmental parameter distribution map. The central fusion module uses the FOP-MOC model.

[0102] This invention employs sensor modules resistant to high humidity, high dust, and extreme temperatures, deployed in different areas of the farm to collect data on temperature, humidity, light intensity, and harmful gas concentrations. An adaptive filtering algorithm is integrated into the sensor modules to preprocess data in high-humidity, high-dust environments, removing noise interference and improving data accuracy. A data caching mechanism is established in the cloud platform; if data packet loss or delay is detected, missing data is automatically resent from the cache to ensure data continuity. Based on the cached data, multiple types of sensors are deployed to address environmental differences in different areas, establishing a regionally differentiated monitoring model in the cloud platform to define environmental standards for each area and achieve refined management. Edge computing technology is used to perform preliminary data processing locally on the sensor nodes, extracting key features before transmitting to the cloud platform, reducing data redundancy. Simultaneously, through equipment health monitoring and data fusion algorithms, a global environmental parameter distribution map is provided. This invention effectively solves the challenges of data collection, transmission, and analysis in complex environments of farms, achieving refined environmental management and improving farming efficiency and safety.

[0103] Example 2

[0104] The present invention also provides an Internet of Things-based livestock production environment monitoring system, the system being used to implement the method described in any one of the embodiments, the system comprising: a preprocessing module, a feature extraction module, an anomaly detection module, and a data fusion module;

[0105] The preprocessing module is used to deploy sensor modules in different areas of the farm to collect data on temperature, humidity, light intensity and harmful gas concentrations and perform data preprocessing.

[0106] The feature extraction module is used to perform preliminary processing on the preprocessed data using edge computing technology, extract key features, and then transmit them to the cloud platform.

[0107] The anomaly detection module is used to establish a regionally differentiated monitoring model in the cloud platform, monitor key characteristic data in real time, and trigger an early warning mechanism if abnormal fluctuations in temperature, humidity or harmful gas concentration are detected, so as to notify the management personnel to handle the situation in a timely manner.

[0108] The data fusion module is used to deploy data fusion algorithms in the cloud platform, integrate sensor data from multiple regions, and generate a global environmental parameter distribution map.

[0109] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for monitoring the livestock production environment based on the Internet of Things, characterized in that, The method includes: Sensor modules were deployed in different areas of the farm to collect data on temperature, humidity, light intensity, and concentration of harmful gases, and the data was preprocessed. Edge computing technology is used to perform preliminary processing on the pre-processed data, extract key features, and then transmit the data to the cloud platform. Establish a regionally differentiated monitoring model in the cloud platform to monitor key characteristic data in real time. If abnormal fluctuations in temperature, humidity or concentration of harmful gases are detected, trigger the early warning mechanism and notify the management personnel to handle the situation in a timely manner. Deploy data fusion algorithms on a cloud platform to integrate sensor data from multiple regions and generate a global environmental parameter distribution map; Methods for establishing regionally differentiated monitoring models in cloud platforms and monitoring key characteristic data in real time include: Establish a conceptual model: In the formula, R, f(H), f(E), f(R), and f(V) represent the regional environmental anomaly, the risk of the anomaly source, the receptor exposure, the receptor resistance, and the receptor vulnerability of the evaluation object, respectively. Each evaluation indicator is dimensionless. For object The j-th indicator in Value of time Then we have: In the formula, The value is the dimensionless value; and The j-th indicator is in Mean and standard deviation at time points; The evaluation of each object was calculated using the vertical and horizontal grading method. The risk index of abnormal sources, receptor exposure, and resilience index at any given time; If abnormal fluctuations in temperature, humidity, or harmful gas concentration are detected, an early warning mechanism will be triggered to notify management personnel for timely handling. Methods for deploying data fusion algorithms on a cloud platform to integrate multi-regional sensor data and generate a global environmental parameter distribution map include: For each region, the K-nearest neighbor classification algorithm is used to match the dimensionality-reduced sensor data; data that closely matches the preset samples in the storage is selected, and then the filtered data is precisely matched by a triplet + neural network with ResNet as the backbone network, thereby obtaining the fused sensor data for each region; All sensor data from each region after fusion are transmitted to the central fusion module, where tracking, correlation, and estimation are performed to generate a global environmental parameter distribution map. The central fusion module uses the FOP-MOC model.

2. The method according to claim 1, characterized in that, Methods for deploying sensor modules in different areas of a farm include: The preprocessed data is transmitted through low-power wide-area network technology. Specifically, relay nodes are deployed in the farm, and the location and number of nodes are planned in a reasonable manner. Establish a data caching mechanism in the cloud platform. If a data packet is detected to be lost or delayed, the missing data will be automatically resent from the cache. Based on cached data, multiple types of sensors are deployed to address environmental differences in different regions.

3. The method according to claim 2, characterized in that, The method of using edge computing technology to perform preliminary processing on preprocessed data and extract key features includes: extracting key features by constructing an edge computing network architecture framework, which includes: service layer, platform layer, transmission layer, edge layer, and perception layer. The perception layer, connected to the edge gateway, can transmit various environmental information to the edge layer in real time. The edge layer mainly consists of edge gateways, edge cloud, and edge controllers. The edge gateway is responsible for managing various sensors and computing resources. The edge cloud provides storage and computing resources for monitoring data, enabling data processing and analysis to be performed locally. The edge controller is responsible for real-time control and optimization of various sensors. The transport layer is primarily responsible for data transmission, selecting the appropriate network for information transmission based on different application scenarios. The platform layer serves as the hub for data collection and processing, analyzing various types of environmental data; The service layer focuses on the analysis of environmental monitoring data. Through in-depth analysis of environmental monitoring data, the service layer can promptly identify abnormal information. The service layer also predicts environmental trends in the future by mining and analyzing historical data. Discrete wavelet transform is introduced to achieve time-frequency domain feature extraction. ； In the formula, Represents the original data. Represents wavelet coefficients, Let denote the wavelet basis function, j denote the scale, and m denote the translation.

4. The method according to claim 3, characterized in that, Methods for establishing regionally differentiated monitoring models in cloud platforms include: An edge computing network architecture is deployed at sensor nodes to acquire environmental data, preprocess the data, and filter out noise. Feature extraction algorithms are used to extract key features from preprocessed data to obtain a feature dataset, which is then transmitted to the cloud platform. The system receives feature data on the cloud platform and uses a data integrity detection algorithm to determine if data is lost during transmission. If data loss is detected, a retransmission mechanism is triggered. Based on the feature dataset, a clustering algorithm is used to classify the environmental data and determine the distribution patterns of environmental features in different regions. Based on the distribution patterns of environmental characteristics, a regional division algorithm is adopted to generate a regionally differentiated environmental monitoring model.

5. A livestock production environment monitoring system based on the Internet of Things, said system being used to implement the method described in any one of claims 1-4, characterized in that, The system includes: a preprocessing module, a feature extraction module, an anomaly detection module, and a data fusion module; The preprocessing module is used to deploy sensor modules in different areas of the farm to collect data on temperature, humidity, light intensity and concentration of harmful gases and perform data preprocessing. The feature extraction module is used to perform preliminary processing on the preprocessed data using edge computing technology, extract key features, and then transmit them to the cloud platform. The anomaly detection module is used to establish a regionally differentiated monitoring model in the cloud platform, monitor key characteristic data in real time, and trigger an early warning mechanism if abnormal fluctuations in temperature, humidity or harmful gas concentration are detected, so as to notify the management personnel to handle the situation in a timely manner. The data fusion module is used to deploy data fusion algorithms in the cloud platform, integrate sensor data from multiple regions, and generate a global environmental parameter distribution map. The process of establishing a regionally differentiated monitoring model in a cloud platform and monitoring key characteristic data in real time includes: Establish a conceptual model: In the formula, R, f(H), f(E), f(R), and f(V) represent the regional environmental anomaly, the risk of the anomaly source, the receptor exposure, the receptor resistance, and the receptor vulnerability of the evaluation object, respectively. Each evaluation indicator is dimensionless. For object The j-th indicator in Value of time Then we have: In the formula, The value is the dimensionless value; and The j-th indicator is in Mean and standard deviation at time points; The evaluation of each object was calculated using the vertical and horizontal grading method. The risk index of abnormal sources, receptor exposure, and resilience index at any given time; If abnormal fluctuations in temperature, humidity, or harmful gas concentration are detected, an early warning mechanism will be triggered to notify management personnel for timely handling. The process of deploying data fusion algorithms on a cloud platform, integrating multi-regional sensor data, and generating a global environmental parameter distribution map includes: For each region, the K-nearest neighbor classification algorithm is used to match the dimensionality-reduced sensor data; data that closely matches the preset samples in the storage is selected, and then the filtered data is precisely matched by a triplet + neural network with ResNet as the backbone network, thereby obtaining the fused sensor data for each region; All sensor data from each region after fusion are transmitted to the central fusion module, where tracking, correlation, and estimation are performed to generate a global environmental parameter distribution map. The central fusion module uses the FOP-MOC model.

Images