Cowshed environment temperature intelligent monitoring system based on wireless sensor network

Abstract

The invention relates to the technical field of agricultural livestock breeding automation equipment, and discloses a cowshed environment temperature intelligent monitoring system based on a wireless sensor network, and the system comprises a thermal imaging recognition module which is provided with an infrared thermal imaging device and is used for obtaining a thermal distribution image in a cowshed space at regular time. The technical scheme that the infrared thermal imaging recognition module is combined with the regional weight calculation and energy perception scheduling mechanism is adopted, and a fusion mechanism of global visual recognition of heat distribution of the internal space of the cowshed and dynamic key region monitoring scheduling is achieved. The intelligent monitoring system for the environment temperature of the cowshed obtains a space heat map through infrared thermal imaging, extracts a heat abnormal area, generates an area weight value in combination with a cowshed structure, further guides a sensing node to intelligently adjust the sampling frequency and the working state according to the weight, and achieves dynamic temperature fine-grained monitoring of high-risk areas of ventilation blind areas and cattle gathering areas.

Description

Technical Field

[0001] The present invention relates to the technical field of agricultural livestock breeding automation equipment, and in particular to an intelligent monitoring system for cattle shed environmental temperature based on a wireless sensor network. Background Art

[0002] Ambient temperature in the cattle barn is the most important environmental factor and the single most important contributor to heat stress. As ambient temperatures rise, beef cattle's body temperatures rise, dry matter intake decreases, and the decline in dry matter intake (DMI) increases with continued high temperatures. Ambient temperature has a lagging effect on body temperature, and body temperature also has a lagging effect on feed intake. After a period of elevated indoor temperatures, tympanic membrane temperature begins to drop, and subsequently, feed intake decreases. As air temperatures rise, the temperature difference between beef cattle and the surrounding environment decreases, making it difficult to dissipate heat, leading to elevated body temperatures. The comfortable temperature range for beef cattle is 5-25°C. When temperatures exceed 26°C, beef cattle experience heat stress.

[0003] For example, the Chinese invention application with publication number CN107494320A discloses an intelligent monitoring system for the ambient temperature of a cowshed based on a wireless sensor network, which is characterized in that the intelligent monitoring system consists of three parts: a cowshed environmental parameter collection and intelligent prediction platform based on a wireless sensor network, a cowshed environmental multi-point temperature fusion model, and an intelligent prediction model for the ambient temperature of a cowshed; the present invention effectively solves the problem that the existing cowshed monitoring system fails to detect and predict the temperature of the cowshed environment based on the nonlinearity and large hysteresis of the ambient temperature changes of the cowshed and the complex temperature changes due to the large area of ​​the cowshed, which greatly affects the monitoring of the ambient temperature of the cowshed.

[0004] The shortcomings of the above patents are:

[0005] On the one hand, existing cattle barn environmental temperature monitoring systems generally lack the ability to identify temperature distribution within local areas of the barn. Because traditional sensors are typically deployed at fixed locations and a constant sampling frequency, they are unable to accurately and timely capture temperature differences that dynamically change over time in microenvironments such as areas where cattle gather and ventilation blind spots. This results in significant blind spots in the system's monitoring, making it impossible to effectively identify potential heat stress risk points, impacting the accuracy and response efficiency of intelligent livestock environmental management.

[0006] On the other hand, in the process of trying to improve system monitoring accuracy, existing technologies often enhance data coverage and timeliness by increasing the number of sensor nodes or increasing the data sampling frequency. However, this results in a significant increase in overall system power consumption, increased load on communication links, and even uneven energy distribution among nodes, which in turn affects the stable operation of the entire network and the reliable transmission of data. Furthermore, high energy consumption and high congestion increase system maintenance costs and limit the long-term deployment and widespread application of monitoring systems on large-scale ranches.

[0007] To this end, the present invention proposes an intelligent monitoring system for cattle house ambient temperature based on a wireless sensor network to solve the above-mentioned problems. Summary of the Invention

[0008] In view of the deficiencies in the prior art, the present invention provides an intelligent monitoring system for the ambient temperature of a cowshed based on a wireless sensor network to solve the problems raised in the above background technology.

[0009] To achieve the above objectives, the present invention is implemented through the following technical solutions: a cowshed environment temperature intelligent monitoring system based on a wireless sensor network, the cowshed environment temperature intelligent monitoring system comprising:

[0010] The thermal imaging recognition module is equipped with an infrared thermal imaging device to regularly acquire thermal distribution images in the cowshed space, identify thermal anomaly areas through image processing, and output thermal anomaly area data;

[0011] The regional weight calculation module receives the output thermal anomaly area data, combines it with the spatial structure information of the cattle shed, and calculates the monitoring priority weight of each node corresponding to the area covered by the wireless sensor node;

[0012] The energy-aware scheduling module comprehensively determines the current working status of the node based on the monitoring priority weight, combined with the remaining battery capacity of each node and the real-time load of the communication link relay task. It also allocates sampling frequency based on the monitoring priority weight, determines whether the node enters a low-power working state, and the frequency of data upload;

[0013] The communication coordination mechanism module receives the working status of the node and the data upload plan, and is used to perform distributed routing control based on the number of hops and load conditions between nodes, thereby achieving priority transmission and load balancing scheduling of data in high-priority areas.

[0014] Preferably, the thermal imaging recognition module is equipped with an infrared thermal imaging device for regularly acquiring thermal distribution images in the cowshed space, further comprising:

[0015] The image preprocessing unit is used to receive the original thermal image data obtained by the infrared thermal imaging device and perform image enhancement and denoising. Image enhancement includes histogram equalization and edge sharpening operations. The denoising process is based on the convolution transformation of the Gaussian filter kernel function. The formula is:

[0016]

[0017] Where G(x,y) is the filter weight value at the coordinate (x,y), σ is the standard deviation of the Gaussian kernel, and x and y are the pixel coordinate offsets;

[0018] The temperature field reconstruction unit is used to restore the temperature value at the pixel level based on the enhanced thermal image output by the image preprocessing unit and the radiation conversion model of the infrared sensor, and convert each pixel in the image into the actual temperature value. The temperature calculation is based on the modified form of the Stefan-Boltzmann law:

[0019]

[0020] Where T(x,y) is the actual temperature value at the coordinate (x,y) in the image, R(x,y) is the infrared radiation intensity value, ε is the emissivity of the cow surface or building surface material, σ B is the Stefan-Boltzmann constant;

[0021] The anomaly detection unit is used to analyze the two-dimensional temperature matrix generated by the temperature field reconstruction unit and detect thermal anomaly areas through temperature gradient analysis and threshold judgment. The calculation of the temperature gradient is based on the neighborhood difference model and is defined as:

[0022]

[0023] If the temperature gradient of the pixel (x,y) If the value is greater than the set threshold θ1, the area is considered to be a temperature drastic change area;

[0024] Further regional connectivity detection is performed on all pixels that meet the conditions to form closed area blocks, and the validity is determined based on whether the temperature variance within the block exceeds the set value θ2:

[0025]

[0026] Among them, Var 区域 is the regional temperature variance, n is the total number of pixels in the region, T i is the temperature value of each pixel, is the regional average temperature;

[0027] The dynamic trend analysis unit is used to perform region matching and state evolution analysis between consecutive frames based on the thermal anomaly region boundaries and position coordinates output by the anomaly detection unit, to determine whether the thermal anomaly region is in a state of continuous expansion, stability, or weakening, and to perform dynamic trend classification using a time series difference model:

[0028] ΔA t =A t -A t-1 ,

[0029] Where ΔA t is the area change of thermal anomaly region, A t is the area of ​​the thermal anomaly region in the tth frame, A t-1is the area of ​​the thermal anomaly region in the t-1th frame, is the domain average temperature change, is the temperature data of the thermal anomaly area in the tth frame, is the temperature data of the thermal anomaly area in the t-1th frame;

[0030] If satisfied:

[0031] ΔA t >δ1 and Determined to be abnormally aggravated;

[0032] ∣ΔA t |<δ3 and Determined to be abnormally stable;

[0033] ΔA t <-δ1 and It was judged to be abnormally weakened;

[0034] Among them, δ1, δ2, δ3, and δ4 are the set area and temperature change thresholds.

[0035] Preferably, the region weight calculation module receives the number, spatial coordinates, temperature feature vector and trend information of the thermal anomaly region output by the thermal imaging recognition module, and further includes:

[0036] The spatial mapping unit is used to align the spatial coordinates of the thermal anomaly area provided by the thermal imaging recognition module with the three-dimensional structural model of the cowshed, identify the spatial overlap between each thermal anomaly area and the wireless sensor node, and calculate the overlap coefficient between the node sensing area and the anomaly area through the spatial coverage model. The calculation formula of the overlap coefficient is:

[0037]

[0038] Among them, λ i,j is the spatial overlap coefficient between the i-th sensing node and the j-th thermal anomaly area, S i,j is the intersection area between the node sensing area and the thermal anomaly area, S j is the total area of ​​the thermal anomaly region;

[0039] When λ i,j When θ > θ1, it is determined that there is a spatial response relationship between the node and the abnormal area, and the relevant node number and spatial coefficient are transferred to the temperature response unit;

[0040] Temperature response unit, used to output the overlap coefficient λ according to the spatial mapping unit i,j , combined with the average temperature value and trend label of the thermal anomaly area, the temperature response intensity of each node is calculated. The temperature response intensity is used to reflect the temperature sensitivity of the node coverage area under the current thermal dynamics. The definition formula is as follows:

[0041]

[0042] Among them, μ i,j is the temperature response intensity of the i-th node to the j-th abnormal area, is the average temperature of the area, T ref The temperature reference value set for the cowshed environment, ψ j is the regional trend weight coefficient, defined as:

[0043] If the trend is abnormally intensified, then ψ j =1.2;

[0044] If the trend is abnormally stable, then ψ j =1.0;

[0045] If the trend is abnormal weakening, then ψ j =0.8;

[0046] The temperature response unit outputs the response intensity value of each node to all abnormal areas in the current frame, which is used for normalization processing by the risk superposition unit.

[0047] Preferably, the region weight calculation module receives the number, spatial coordinates, temperature feature vector and trend information of the thermal anomaly region output by the thermal imaging recognition module, and further includes:

[0048] The risk superposition unit is used to perform node-level aggregation on the multi-region response strength calculated by the temperature response unit, and obtain the total risk perception value of each node through weighted summation. At the same time, the risk weight is set considering the importance difference between regions. The calculation formula is:

[0049]

[0050] Among them, ρ i The total risk response value of the i-th sensor node, μ i,j is the temperature response intensity of the i-th node to the j-th abnormal area, ω j is the importance weight of the jth thermal anomaly area, and the weight is set according to the distance between the area center and the cattle gathering interval:

[0051]

[0052] Among them, d j D is the distance from the center of the thermal anomaly area to the center of gravity of the cattle herd, max is the maximum distance from any point in the cowshed to the center of gravity;

[0053] The risk superposition unit outputs the node-level risk response value, which serves as the input of the final priority scoring unit;

[0054] Priority scoring unit, used to add the node risk response value ρ calculated by the risk superposition unit i Converted to monitoring priority weight w i , combined with the node's own historical response frequency to adjust, to prevent a single high-frequency node from being overloaded for a long time. The specific scoring function is defined as follows:

[0055]

[0056] Among them, w i is the final monitoring priority weight of the i-th node, ρ i The total risk response value of the i-th sensor node, ρ max is the maximum risk response value of all nodes, f i is the frequency of scheduling the node within 1 day, f max The maximum scheduling frequency for all nodes;

[0057] The priority weight w output by the priority scoring unit i It will be passed to the energy-aware scheduling module as one of the node scheduling bases.

[0058] Preferably, in the energy-aware scheduling module, the node monitoring priority weight w output by the receiving area weight calculation module is i , combined with the current remaining power E of each node i and the relay task load L in the communication link i , further including:

[0059] The state determination unit is used to determine the node priority weight w i 、Remaining power E i and load factor L i , comprehensively evaluate whether each node is currently in an activated state;

[0060] The state judgment unit adopts the energy load joint judgment function:

[0061]

[0062] Among them, S i Is node i activated? max The node battery is fully charged, L max is the maximum link load value in the system, θ s is the state activation threshold;

[0063] Sampling scheduling unit, used to activate node S i =1 is scheduled, and the sampling frequency is based on the monitoring priority weight w i Determined jointly with the node power factor; the sampling frequency is set to:

[0064]

[0065] in, The current sampling frequency of the i-th node, f base is the basic sampling frequency, β1 is the monitoring weight coefficient, and β2 is the power regulation coefficient;

[0066] The sampling frequency is limited to a maximum value f max If the calculation result exceeds the limit, truncation is performed.

[0067] Preferably, in the energy-aware scheduling module, the node monitoring priority weight w output by the receiving area weight calculation module is i , combined with the current remaining power E of each node i and the relay task load L in the communication link i , further including:

[0068] The upload scheduling unit is used to set the node data upload cycle. The upload cycle is based on the sampling frequency and the node's network load L. i Joint control is defined as:

[0069]

[0070] in, is the node upload interval, γ is the load adjustment coefficient;

[0071] Power consumption control unit, used to calculate the sampling frequency of the node Upload cycle Remaining power E i Calculate the power consumption risk index and determine whether the node needs to enter low-power working mode. The power consumption risk index is defined as:

[0072]

[0073] Among them, R i is the power consumption risk index, is the upload cycle;

[0074] If R i ≥θ r , where θ r The node is put into low power mode and the polling wake-up mechanism is enabled when the node reaches the power consumption risk threshold.

[0075] Preferably, the communication coordination mechanism module further includes:

[0076] Topology perception unit, used to build the current network topology map and record the number of hops h from each node to the sink node i , current forwarding load ξ i , and dynamically maintain the neighbor node set N i, where the hop count h i Obtained through broadcast network initialization and periodically updated after topology changes;

[0077] Define the forwarding load index:

[0078]

[0079] Among them, M k is the predicted value of the number of uploaded data packets of neighbor node k, is the upload cycle, ξ i is the relay load pressure at the location of node i;

[0080] The topology perception unit outputs the topology hop count h of each node i and load index ξ i , for use by subsequent routing path evaluation units;

[0081] The routing path evaluation unit is used to comprehensively evaluate the forwarding priority value of each feasible path, with the goal of selecting a path with low load, short distance and support for high priority weight node data;

[0082] For each neighbor j∈N of node i i , define the routing forwarding score function:

[0083]

[0084] Among them, φ i,j is the score of forwarding data through neighbor node j, h j is the number of hops from node j to the sink node, ξ j is the load index of node j, w j is the priority weight, S j is in the activated state;

[0085] The routing path evaluation unit outputs a score list of all available forwarding paths {φ i,j}, provided to the priority transmission scheduling unit for path selection.

[0086] Preferably, the communication coordination mechanism module further includes:

[0087] The priority transmission scheduling unit is used to select the optimal forwarding path for data of different priorities based on the routing path scoring results; each node in the transmission queue is assigned the node priority weight w corresponding to the data. i Sort tasks and set scheduling factors:

[0088]

[0089] Among them, η iis the importance data density per unit time of node i, and the scheduling priority queue is based on η i Sort by value in descending order;

[0090] For each data to be forwarded, the path score φ calculated in the previous unit i,j Select the activated neighbor node j with the highest score * As forwarding target:

[0091]

[0092] At the same time, avoid long-term overload of a single node. Reselect the node with the next largest φ from the remaining candidate set;

[0093] Congestion feedback adjustment unit, used to adjust routing and rate control when buffer backlog or forwarding failure occurs at the node; monitoring indicators include: number of consecutive failures F i , Current cache usage B i , average waiting time D i , set the adjustment function:

[0094]

[0095] If K i ≥θ c , determine that the node enters the congested state, and immediately perform the following operations:

[0096] Reduce local forwarding frequency

[0097] Notify the upstream node to temporarily disable this node as a forwarding path;

[0098] The feedback scheduling module re-evaluates the forwarding path score φ i,j , remove the congested nodes and reselect j * ;

[0099] The congestion feedback regulation unit outputs a congestion feedback flag and a path adjustment notification to achieve adaptive stability control of the system.

[0100] Preferably, the intelligent monitoring system for the ambient temperature of the cowshed receives sensor reported data through an edge computing node and performs local preprocessing, and the edge node uploads the processed key data to a remote server to reduce communication bandwidth consumption.

[0101] Preferably, the intelligent monitoring system for the ambient temperature of the cowshed can realize iterative optimization of the monitoring strategy based on the dynamic changes of the hot spot area. The system updates the node layout plan and energy scheduling strategy within a preset time period to adapt to the environmental evolution characteristics.

[0102] The present invention provides an intelligent monitoring system for cattle house ambient temperature based on a wireless sensor network. It has the following beneficial effects:

[0103] 1. The present invention adopts a technical solution that combines an infrared thermal imaging recognition module with regional weight calculation and an energy perception scheduling mechanism to realize a fusion mechanism of global visual recognition of the internal thermal distribution of the cowshed and dynamic key area monitoring and scheduling. The intelligent monitoring system for the environmental temperature of the cowshed obtains a spatial thermal map through infrared thermal imaging, extracts thermal anomaly areas, generates regional weight values ​​in combination with the structure of the cowshed, and then guides the sensor nodes to intelligently adjust the sampling frequency and working status according to the weight, so as to achieve dynamic temperature fine-grained monitoring of high-risk areas such as ventilation blind spots and cattle gathering areas. Compared with the static layout scheme of the prior art that adopts fixed points and constant frequency sampling, which cannot accurately identify thermal anomaly risk areas and respond in real time, the present invention significantly improves the recognition ability and response timeliness of the intelligent monitoring system for the environmental temperature of the cowshed to the temperature differences of the local microenvironment, and effectively solves the deficiency of heat stress warning failure caused by monitoring blind spots.

[0104] 2. The present invention adopts a distributed energy-saving optimization solution based on energy-aware scheduling and communication coordination mechanism. Through state judgment, sampling frequency control, upload cycle planning and path optimization mechanism, it dynamically adjusts node energy consumption and communication pressure while ensuring monitoring coverage and data timeliness, and constructs an adaptive network control system for load balancing and energy consumption balance. Compared with the high-power consumption solution adopted in the prior art to improve monitoring accuracy by blindly increasing the number of nodes or increasing the overall sampling frequency, which leads to a rapid increase in system energy consumption and frequent communication bottlenecks, the present invention realizes efficient information transmission under low power consumption conditions, solves the bottlenecks of uncontrollable energy consumption, uneven load distribution and poor network stability of the existing system, and is suitable for long-term deployment in large-scale ranches. BRIEF DESCRIPTION OF THE DRAWINGS

[0105] Figure 1 It is a system diagram of the present invention. DETAILED DESCRIPTION

[0106] To help those skilled in the art understand the present invention, the following will provide a clear and complete description of the technical solutions in the embodiments of the present invention, in conjunction with the accompanying drawings. Obviously, the described embodiments are only partial embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0107] The present invention is described in detail below with reference to the accompanying drawings:

[0108] Example:

[0109] Please see the attached Figure 1The embodiment of the present invention provides a cowshed environment temperature intelligent monitoring system based on a wireless sensor network, and the cowshed environment temperature intelligent monitoring system includes:

[0110] The thermal imaging recognition module is equipped with an infrared thermal imaging device to regularly acquire thermal distribution images in the cowshed space, identify thermal anomaly areas through image processing, and output thermal anomaly area data;

[0111] The image preprocessing unit is used to receive the original thermal image data obtained by the infrared thermal imaging device and perform image enhancement and denoising. Image enhancement includes histogram equalization and edge sharpening operations. The denoising process is based on the convolution transformation of the Gaussian filter kernel function. The formula is:

[0112]

[0113] Where G(x,y) is the filter weight value at the coordinate (x,y), σ is the standard deviation of the Gaussian kernel, and x and y are the pixel coordinate offsets;

[0114] The temperature field reconstruction unit is used to restore the temperature value at the pixel level based on the enhanced thermal image output by the image preprocessing unit and the radiation conversion model of the infrared sensor, and convert each pixel in the image into the actual temperature value. The temperature calculation is based on the modified form of the Stefan-Boltzmann law:

[0115]

[0116] Where T(x,y) is the actual temperature value at the coordinate (x,y) in the image, R(x,y) is the infrared radiation intensity value, ε is the emissivity of the cow surface or building surface material, σ B is the Stefan-Boltzmann constant;

[0117] The anomaly detection unit is used to analyze the two-dimensional temperature matrix generated by the temperature field reconstruction unit and detect thermal anomaly areas through temperature gradient analysis and threshold judgment. The calculation of the temperature gradient is based on the neighborhood difference model and is defined as:

[0118]

[0119] If the temperature gradient of the pixel (x,y) If the value is greater than the set threshold θ1, the area is considered to be a temperature drastic change area;

[0120] Further regional connectivity detection is performed on all pixels that meet the conditions to form closed area blocks, and the validity is determined based on whether the temperature variance within the block exceeds the set value θ2:

[0121]

[0122] Among them, Var区域 is the regional temperature variance, n is the total number of pixels in the region, T i is the temperature value of each pixel, is the regional average temperature;

[0123] The dynamic trend analysis unit is used to perform region matching and state evolution analysis between consecutive frames based on the thermal anomaly region boundaries and position coordinates output by the anomaly detection unit, to determine whether the thermal anomaly region is in a state of continuous expansion, stability, or weakening, and to perform dynamic trend classification using a time series difference model:

[0124] ΔA t =A t -A t-1 ,

[0125] Where ΔA t is the area change of thermal anomaly region, A t is the area of ​​the thermal anomaly region in the tth frame, A t-1 is the area of ​​the thermal anomaly region in the t-1th frame, is the domain average temperature change, is the temperature data of the thermal anomaly area in the tth frame, is the temperature data of the thermal anomaly area in the t-1th frame;

[0126] If satisfied:

[0127] ΔA t >δ1 and Determined to be abnormally aggravated;

[0128] ∣ΔA t |<δ3 and Determined to be abnormally stable;

[0129] ΔA t <-δ1 and It was judged to be abnormally weakened;

[0130] Among them, δ1, δ2, δ3, and δ4 are the set area and temperature change thresholds;

[0131] The regional weight calculation module receives the output thermal anomaly area data, combines it with the spatial structure information of the cattle shed, and calculates the monitoring priority weight of each node corresponding to the area covered by the wireless sensor node;

[0132] The spatial mapping unit is used to align the spatial coordinates of the thermal anomaly area provided by the thermal imaging recognition module with the three-dimensional structural model of the cowshed, identify the spatial overlap between each thermal anomaly area and the wireless sensor node, and calculate the overlap coefficient between the node sensing area and the anomaly area through the spatial coverage model. The calculation formula of the overlap coefficient is:

[0133]

[0134] Among them, λ i,j is the spatial overlap coefficient between the i-th sensing node and the j-th thermal anomaly area, S i,j is the intersection area between the node sensing area and the thermal anomaly area, S j is the total area of ​​the thermal anomaly region;

[0135] When λ i,j When θ > θ1, it is determined that there is a spatial response relationship between the node and the abnormal area, and the relevant node number and spatial coefficient are transferred to the temperature response unit;

[0136] Temperature response unit, used to output the overlap coefficient λ according to the spatial mapping unit i,j , combined with the average temperature value and trend label of the thermal anomaly area, the temperature response intensity of each node is calculated. The temperature response intensity is used to reflect the temperature sensitivity of the node coverage area under the current thermal dynamics. The definition formula is as follows:

[0137]

[0138] Among them, μ i,j is the temperature response intensity of the i-th node to the j-th abnormal area, is the average temperature of the area, T ref The temperature reference value set for the cowshed environment, ψ j is the regional trend weight coefficient, defined as:

[0139] If the trend is abnormally intensified, then ψ j =1.2;

[0140] If the trend is abnormally stable, then ψ j =1.0;

[0141] If the trend is abnormal weakening, then ψ j =0.8;

[0142] The temperature response unit outputs the response intensity value of each node to all abnormal areas in the current frame, which is used for normalization processing by the risk superposition unit;

[0143] The risk superposition unit is used to perform node-level aggregation on the multi-region response strength calculated by the temperature response unit, and obtain the total risk perception value of each node through weighted summation. At the same time, the risk weight is set considering the importance difference between regions. The calculation formula is:

[0144]

[0145] Among them, ρ i The total risk response value of the i-th sensor node, μi,j is the temperature response intensity of the i-th node to the j-th abnormal area, ω j is the importance weight of the jth thermal anomaly area, and the weight is set according to the distance between the area center and the cattle gathering interval:

[0146]

[0147] Among them, d j D is the distance from the center of the thermal anomaly area to the center of gravity of the cattle herd, max is the maximum distance from any point in the cowshed to the center of gravity;

[0148] The risk superposition unit outputs the node-level risk response value, which serves as the input of the final priority scoring unit;

[0149] Priority scoring unit, used to add the node risk response value ρ calculated by the risk superposition unit i Converted to monitoring priority weight w i , combined with the node's own historical response frequency to adjust, to prevent a single high-frequency node from being overloaded for a long time. The specific scoring function is defined as follows:

[0150]

[0151] Among them, w i is the final monitoring priority weight of the i-th node, ρ i The total risk response value of the i-th sensor node, ρ max is the maximum risk response value of all nodes, f i is the frequency of scheduling the node within 1 day, f max The maximum scheduling frequency for all nodes;

[0152] The priority weight w output by the priority scoring unit i The data will be passed to the energy-aware scheduling module as one of the node scheduling bases;

[0153] The energy-aware scheduling module comprehensively determines the current working status of the node based on the monitoring priority weight, combined with the remaining battery capacity of each node and the real-time load of the communication link relay task. It also allocates sampling frequency based on the monitoring priority weight, determines whether the node enters a low-power working state, and the frequency of data upload;

[0154] The state determination unit is used to determine the node priority weight w i 、Remaining power E i and load factor L i , comprehensively evaluate whether each node is currently in an activated state;

[0155] The state judgment unit adopts the energy load joint judgment function:

[0156]

[0157] Among them, S i Is node i activated? max The node battery is fully charged, L max is the maximum link load value in the system, θ s is the state activation threshold;

[0158] Sampling scheduling unit, used to activate node S i =1 is scheduled, and the sampling frequency is based on the monitoring priority weight w i Determined jointly with the node power factor; the sampling frequency is set to:

[0159]

[0160] in, The current sampling frequency of the i-th node, f base is the basic sampling frequency, β1 is the monitoring weight coefficient, and β2 is the power regulation coefficient;

[0161] The sampling frequency is limited to a maximum value f max If the calculation result exceeds the limit, truncation is performed;

[0162] The upload scheduling unit is used to set the node data upload cycle. The upload cycle is based on the sampling frequency and the node's network load L. i Joint control is defined as:

[0163]

[0164] in, is the node upload interval, γ is the load adjustment coefficient;

[0165] Power consumption control unit, used to calculate the sampling frequency of the node Upload cycle Remaining power E i Calculate the power consumption risk index and determine whether the node needs to enter low-power working mode. The power consumption risk index is defined as:

[0166]

[0167] Among them, R i is the power consumption risk index, is the upload cycle;

[0168] If R i ≥θ r , where θ r For the power consumption risk threshold, the node is put into low power consumption mode and the polling wake-up mechanism is enabled;

[0169] The communication coordination mechanism module receives the node working status and data upload plan, and is used to perform distributed routing control based on the number of hops and load conditions between nodes, thereby achieving priority transmission and load balancing scheduling of high-priority area data;

[0170] Topology perception unit, used to build the current network topology map and record the number of hops h from each node to the sink node i , current forwarding load ξ i , and dynamically maintain the neighbor node set N i , where the hop count h i Obtained through broadcast network initialization and periodically updated after topology changes;

[0171] Define the forwarding load index:

[0172]

[0173] Among them, M k is the predicted value of the number of uploaded data packets of neighbor node k, is the upload cycle, ξ i is the relay load pressure at the location of node i;

[0174] The topology perception unit outputs the topology hop count h of each node i and load index ξ i , for use by subsequent routing path evaluation units;

[0175] The routing path evaluation unit is used to comprehensively evaluate the forwarding priority value of each feasible path, with the goal of selecting a path with low load, short distance and support for high priority weight node data;

[0176] For each neighbor j∈N of node i i , define the routing forwarding score function:

[0177]

[0178] Among them, φ i,j is the score of forwarding data through neighbor node j, h j is the number of hops from node j to the sink node, ξ j is the load index of node j, w j is the priority weight, S j is in the activated state;

[0179] The routing path evaluation unit outputs a score list of all available forwarding paths {φ i,j}, provided to the priority transmission scheduling unit for path selection;

[0180] The priority transmission scheduling unit is used to select the optimal forwarding path for data of different priorities based on the routing path scoring results; each node in the transmission queue is assigned the node priority weight w corresponding to the data. i Sort tasks and set scheduling factors:

[0181]

[0182] Among them, η i is the importance data density per unit time of node i, and the scheduling priority queue is based on η i Sort by value in descending order;

[0183] For each data to be forwarded, the path score φ calculated in the previous unit i,j Select the activated neighbor node j with the highest score * As forwarding target:

[0184]

[0185] At the same time, avoid long-term overload of a single node. Reselect the node with the next largest φ from the remaining candidate set;

[0186] Congestion feedback adjustment unit, used to adjust routing and rate control when buffer backlog or forwarding failure occurs at the node; monitoring indicators include: number of consecutive failures F i , Current cache usage B i , average waiting time D i , set the adjustment function:

[0187]

[0188] If K i ≥θ c , determine that the node enters the congested state, and immediately perform the following operations:

[0189] Reduce local forwarding frequency

[0190] Notify the upstream node to temporarily disable this node as a forwarding path;

[0191] The feedback scheduling module re-evaluates the forwarding path score φ i,j , remove the congested nodes and reselect j * ;

[0192] The congestion feedback regulation unit outputs a congestion feedback flag and a path adjustment notification to achieve adaptive stability control of the system.

[0193] The benefits of the thermal imaging recognition module include: It uses infrared equipment to achieve panoramic perception of the heat distribution throughout the entire cattle barn space. Its non-contact, high-resolution, and wide-area coverage advantages allow for real-time monitoring of temperature changes in various areas within the barn without disrupting cattle movement. Compared to traditional single-point temperature collection methods, the thermal imaging recognition module significantly enhances the system's ability to identify spatial heterogeneity in ambient temperature, particularly in key areas where cattle gather and ventilation blind spots, enabling rapid identification and detailed local analysis, providing precise guidance for subsequent scheduling and early warning.

[0194] The benefits of the regional weight calculation module include analyzing the spatial overlap between thermal anomaly areas and sensor nodes, combining regional temperature trends with the importance of structural locations, and dynamically assigning weights to each node, enabling precise resource allocation. The regional weight calculation module enables regional sensitivity and dynamic adaptability in data scheduling, effectively avoiding resource waste and low-value information collection, and strengthening the system's ability to perceive and focus on critical environmental areas.

[0195] The Energy-Aware Scheduling Module offers benefits. By combining node priority, power status, and network load, it effectively implements intelligent control over node activation status, sampling frequency, and data upload cycle, ensuring optimal monitoring efficiency within limited energy resources. This module significantly extends node lifespan, reduces system maintenance frequency, and ensures uncompromised monitoring accuracy in high-priority areas, making it a key component in achieving a balance between energy conservation and performance.

[0196] The benefits of the communication coordination mechanism module include: It implements distributed routing control based on node topology, load status, and priority weights, ensuring that critical data is rapidly transmitted along the optimal path. It also dynamically adjusts transmission strategies through congestion awareness and feedback regulation mechanisms to prevent link congestion and data loss. The communication coordination mechanism module significantly improves system stability and reliability in high-load, multi-hop communication environments, and serves as the core communication assurance unit supporting the large-scale deployment and high-availability operation of intelligent perception systems.

[0197] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent monitoring system for cattle house ambient temperature based on wireless sensor network, characterized in that: The intelligent monitoring system for the ambient temperature of the cowshed comprises: The thermal imaging recognition module is equipped with an infrared thermal imaging device to regularly acquire thermal distribution images in the cowshed space, identify thermal anomaly areas through image processing, and output thermal anomaly area data; The regional weight calculation module receives the output thermal anomaly area data, combines it with the spatial structure information of the cattle shed, and calculates the monitoring priority weight of each node corresponding to the area covered by the wireless sensor node; The energy-aware scheduling module comprehensively determines the current working status of the node based on the monitoring priority weight, combined with the remaining battery capacity of each node and the real-time load of the communication link relay task. It also allocates sampling frequency based on the monitoring priority weight, determines whether the node enters a low-power working state, and the frequency of data upload; The communication coordination mechanism module receives the working status of the node and the data upload plan, and is used to perform distributed routing control based on the number of hops and load conditions between nodes, thereby achieving priority transmission and load balancing scheduling of data in high-priority areas.

2. The intelligent monitoring system for cattle house ambient temperature based on wireless sensor network according to claim 1, characterized in that: The thermal imaging recognition module is equipped with an infrared thermal imaging device for regularly acquiring thermal distribution images in the cowshed space, and further includes: The image preprocessing unit is used to receive the original thermal image data obtained by the infrared thermal imaging device and perform image enhancement and denoising. Image enhancement includes histogram equalization and edge sharpening operations. The denoising process is based on the convolution transformation of the Gaussian filter kernel function. The formula is: Where G(x,y) is the filter weight value at the coordinate (x,y), σ is the standard deviation of the Gaussian kernel, and x and y are the pixel coordinate offsets; The temperature field reconstruction unit is used to restore the temperature value at the pixel level based on the enhanced thermal image output by the image preprocessing unit and the radiation conversion model of the infrared sensor, and convert each pixel in the image into the actual temperature value. The temperature calculation is based on the modified form of the Stefan-Boltzmann law: Where T(x,y) is the actual temperature value at the coordinate (x,y) in the image, R(x,y) is the infrared radiation intensity value, ε is the emissivity of the cow surface or building surface material, σ B is the Stefan-Boltzmann constant; The anomaly detection unit is used to analyze the two-dimensional temperature matrix generated by the temperature field reconstruction unit and detect thermal anomaly areas through temperature gradient analysis and threshold judgment. The calculation of the temperature gradient is based on the neighborhood difference model and is defined as: If the temperature gradient of the pixel (x,y) If the value is greater than the set threshold θ1, the area is considered to be a temperature drastic change area; Further regional connectivity detection is performed on all pixels that meet the conditions to form closed area blocks, and the validity is determined based on whether the temperature variance within the block exceeds the set value θ2: Among them, Var 区域 is the regional temperature variance, n is the total number of pixels in the region, T i is the temperature value of each pixel, T is the average temperature of the region; The dynamic trend analysis unit is used to perform region matching and state evolution analysis between consecutive frames based on the thermal anomaly region boundaries and position coordinates output by the anomaly detection unit, to determine whether the thermal anomaly region is in a state of continuous expansion, stability, or weakening, and to perform dynamic trend classification using a time series difference model: ΔA t = Yes t -IN t-1 , Where ΔA t is the area change of thermal anomaly region, A t is the area of ​​the thermal anomaly region in the tth frame, A t-1 is the area of ​​the thermal anomaly region in the t-1th frame, is the domain average temperature change, is the temperature data of the thermal anomaly area in the tth frame, is the temperature data of the thermal anomaly area in the t-1th frame; If satisfied: ΔA t >δ1 and Determined to be abnormally aggravated; ∣ΔA t |<δ3 and Determined to be abnormally stable; ΔA t <-δ1 and It was judged to be abnormally weakened; Among them, δ1, δ2, δ3, and δ4 are the set area and temperature change thresholds.

3. The intelligent monitoring system for cowshed ambient temperature based on wireless sensor network according to claim 1, characterized in that: The region weight calculation module receives the number, spatial coordinates, temperature feature vector and trend information of the thermal anomaly region output by the thermal imaging recognition module, and further includes: The spatial mapping unit is used to align the spatial coordinates of the thermal anomaly area provided by the thermal imaging recognition module with the three-dimensional structural model of the cowshed, identify the spatial overlap between each thermal anomaly area and the wireless sensor node, and calculate the overlap coefficient between the node sensing area and the anomaly area through the spatial coverage model. The calculation formula of the overlap coefficient is: Among them, λ i,j is the spatial overlap coefficient between the i-th sensing node and the j-th thermal anomaly area, S i,j is the intersection area between the node sensing area and the thermal anomaly area, S j is the total area of ​​the thermal anomaly region; When λ i,j When θ > θ1, it is determined that there is a spatial response relationship between the node and the abnormal area, and the relevant node number and spatial coefficient are transferred to the temperature response unit; Temperature response unit, used to output the overlap coefficient λ according to the spatial mapping unit i,j , combined with the average temperature value and trend label of the thermal anomaly area, the temperature response intensity of each node is calculated. The temperature response intensity is used to reflect the temperature sensitivity of the node coverage area under the current thermal dynamics. The definition formula is as follows: Among them, μ i,j is the temperature response intensity of the i-th node to the j-th abnormal area, is the average temperature of the area, T ref The temperature reference value set for the cowshed environment, ψ j is the regional trend weight coefficient, defined as: If the trend is abnormally intensified, then ψ j =1.2; If the trend is abnormally stable, then ψ j =1.0; If the trend is abnormal weakening, then ψ j =0.8; The temperature response unit outputs the response intensity value of each node to all abnormal areas in the current frame, which is used for normalization processing by the risk superposition unit.

4. The intelligent monitoring system for cattle house ambient temperature based on wireless sensor network according to claim 1, characterized in that: The region weight calculation module receives the number, spatial coordinates, temperature feature vector and trend information of the thermal anomaly region output by the thermal imaging recognition module, and further includes: The risk superposition unit is used to perform node-level aggregation on the multi-region response strength calculated by the temperature response unit, and obtain the total risk perception value of each node through weighted summation. At the same time, the risk weight is set considering the importance difference between regions. The calculation formula is: Among them, ρ i The total risk response value of the i-th sensor node, μ i,j is the temperature response intensity of the i-th node to the j-th abnormal area, ω j is the importance weight of the jth thermal anomaly area, and the weight is set according to the distance between the area center and the cattle gathering interval: Among them, d j D is the distance from the center of the thermal anomaly area to the center of gravity of the cattle herd, max is the maximum distance from any point in the cowshed to the center of gravity; The risk superposition unit outputs the node-level risk response value, which serves as the input of the final priority scoring unit; Priority scoring unit, used to add the node risk response value ρ calculated by the risk superposition unit i Converted to monitoring priority weight w i , combined with the node's own historical response frequency to adjust, to prevent a single high-frequency node from being overloaded for a long time. The specific scoring function is defined as follows: Among them, w i is the final monitoring priority weight of the i-th node, ρ i The total risk response value of the i-th sensor node, ρ max is the maximum risk response value of all nodes, f i is the frequency of scheduling the node within 1 day, f max The maximum scheduling frequency for all nodes; The priority weight w output by the priority scoring unit i It will be passed to the energy-aware scheduling module as one of the node scheduling bases.

5. The intelligent monitoring system for cattle house ambient temperature based on wireless sensor network according to claim 1, characterized in that: In the energy-aware scheduling module, the node monitoring priority weight w output by the receiving area weight calculation module is i , combined with the current remaining power E of each node i and the relay task load L in the communication link i , further including: The state determination unit is used to determine the node priority weight w i 、Remaining power E i and load factor L i , comprehensively evaluate whether each node is currently in an activated state; The state judgment unit adopts the energy load joint judgment function: Among them, S i Is node i activated? max The node battery is fully charged, L max is the maximum link load value in the system, θ s is the state activation threshold; Sampling scheduling unit, used to activate node S i =1 is scheduled, and the sampling frequency is based on the monitoring priority weight w i Determined jointly with the node power factor; the sampling frequency is set to: in, The current sampling frequency of the i-th node, f base is the basic sampling frequency, β1 is the monitoring weight coefficient, and β2 is the power regulation coefficient; The sampling frequency is limited to a maximum value f max If the calculation result exceeds the limit, truncation is performed.

6. The intelligent monitoring system for cattle house ambient temperature based on wireless sensor network according to claim 1, characterized in that: In the energy-aware scheduling module, the node monitoring priority weight w output by the receiving area weight calculation module is i , combined with the current remaining power E of each node i and the relay task load L in the communication link i , further including: The upload scheduling unit is used to set the node data upload cycle. The upload cycle is based on the sampling frequency and the node's network load L. i Joint control is defined as: in, is the node upload interval, γ is the load adjustment coefficient; Power consumption control unit, used to calculate the sampling frequency of the node Upload cycle Remaining power E i Calculate the power consumption risk index and determine whether the node needs to enter low-power working mode. The power consumption risk index is defined as: Among them, R i is the power consumption risk index, is the upload cycle; If R i ≥θ r , where θ r The node is put into low power mode and the polling wake-up mechanism is enabled when the node reaches the power consumption risk threshold.

7. The intelligent monitoring system for cattle house ambient temperature based on wireless sensor network according to claim 1, characterized in that: The communication coordination mechanism module further includes: Topology perception unit, used to build the current network topology map and record the number of hops h from each node to the sink node i , current forwarding load ξ i , and dynamically maintain the neighbor node set N i , where the hop count h i Obtained through broadcast network initialization and periodically updated after topology changes; Define the forwarding load index: Among them, M k is the predicted value of the number of uploaded data packets of neighbor node k, is the upload cycle, ξ i is the relay load pressure at the location of node i; The topology perception unit outputs the topology hop count h of each node i and load index ξ i , for use by subsequent routing path evaluation units; The routing path evaluation unit is used to comprehensively evaluate the forwarding priority value of each feasible path, with the goal of selecting a path with low load, short distance and support for high priority weight node data; For each neighbor j∈N of node i i , define the routing forwarding score function: Among them, φ i,j is the score of forwarding data through neighbor node j, h j is the number of hops from node j to the sink node, ξ j is the load index of node j, w j is the priority weight, S j is in the activated state; The routing path evaluation unit outputs a score list of all available forwarding paths {φ i,j }, provided to the priority transmission scheduling unit for path selection.

8. The intelligent monitoring system for cattle house ambient temperature based on wireless sensor network according to claim 1, characterized in that: The communication coordination mechanism module further includes: The priority transmission scheduling unit is used to select the optimal forwarding path for data of different priorities based on the routing path scoring results; each node in the transmission queue is assigned the node priority weight w corresponding to the data. i Sort tasks and set scheduling factors: Among them, η i is the importance data density per unit time of node i, and the scheduling priority queue is based on η i Sort by value in descending order; For each data to be forwarded, the path score φ calculated in the previous unit i,j Select the activated neighbor node j with the highest score * As forwarding target: At the same time, avoid long-term overload of a single node. Reselect the node with the next largest φ from the remaining candidate set; Congestion feedback adjustment unit, used to adjust routing and rate control when buffer backlog or forwarding failure occurs at the node; monitoring indicators include: number of consecutive failures F i , Current cache usage B i , average waiting time D i , set the adjustment function: If K i ≥θ c , determine that the node enters the congested state, and immediately perform the following operations: Reduce local forwarding frequency Notify the upstream node to temporarily disable this node as a forwarding path; The feedback scheduling module re-evaluates the forwarding path score φ i,j , remove the congested nodes and reselect j * ; The congestion feedback regulation unit outputs a congestion feedback flag and a path adjustment notification to achieve adaptive stability control of the system.

9. The intelligent monitoring system for cattle house ambient temperature based on wireless sensor network according to claim 1, characterized in that: The intelligent monitoring system for the ambient temperature of the cowshed receives sensor-reported data through edge computing nodes and performs local preprocessing. The edge nodes upload the processed key data to a remote server to reduce communication bandwidth consumption.

10. The intelligent monitoring system for cattle house ambient temperature based on wireless sensor network according to claim 1, characterized in that: The intelligent monitoring system for the ambient temperature of the cowshed can realize iterative optimization of the monitoring strategy based on the dynamic changes of the hot spot area. The system updates the node layout plan and energy scheduling strategy within a preset time period to adapt to the environmental evolution characteristics.

Images