Under-forest chicken raising microclimate self-adaptive regulation method based on temperature sensor

Abstract

This invention relates to the field of temperature sensor testing, specifically the application of testing and measurement technology in the control of microclimates in aquaculture. The method involves deploying a sealed reference cavity with a built-in standard thermometer in a forest-based aquaculture area. Each temperature sensor node is periodically inserted into the reference cavity to periodically read the current temperature value and compare it with the standard thermometer to obtain the absolute drift. A two-dimensional drift matrix is ​​constructed by combining this data with historical humidity data of the location. During non-calibration periods, the original temperature values ​​are corrected using the two-dimensional drift matrix. The corrected temperature values ​​are then used in a Kriging space interpolation algorithm to generate a three-dimensional temperature field distribution map. Based on the coordinates of local high-temperature areas in the distribution map, the corresponding spray cooling solenoid valves are activated. This invention offsets the long-term drift caused by the high-humidity and corrosive environment of the forest undergrowth, corrects the spatial deviation of the original temperature data, improves the accuracy of the three-dimensional temperature field reconstruction, and avoids control misalignment caused by temperature field assessment distortion.

Description

Technical Field

[0001] This invention relates to the field of temperature sensor testing, specifically to the application of testing and measurement technology in the control of microclimate in aquaculture, and discloses an adaptive control method for microclimate in understory chicken farming based on temperature sensors. Background Technology

[0002] Microclimate control in forest-based chicken farming relies on a distributed temperature sensor network. The most readily available approach typically involves a grid-like arrangement of multiple temperature sensor probes within the farming area, with a one-time temperature and electrical signal calibration performed under standard conditions at the factory. During system operation, the main control unit directly collects real-time raw temperature data from each probe and reconstructs the spatial temperature field distribution of the forest-based farming area using spatial interpolation algorithms. When a localized high-temperature region appears in the reconstructed temperature field and its temperature exceeds a preset threshold, the system sends a control command to the coordinates of this high-temperature region, activating the corresponding spray cooling solenoid valve for localized cooling intervention.

[0003] The forest-based aquaculture environment is characterized by persistently high humidity, and the fermentation of chicken manure produces corrosive gases. Temperature sensor probes, exposed to this environment for extended periods, experience surface aging or micro-corrosion, leading to irreversible long-term drift in the temperature-resistance characteristic curves of the sensors. Existing technologies only perform a one-time calibration at the factory, lacking the means to perform in-situ calibration of sensors in complex field environments during system operation. The main control equipment continuously reconstructs the spatial temperature field using the drifted original temperature data, resulting in distortion of the reconstructed temperature field. The system misinterprets the temperature data deviation caused by the drift as an environmental heat island, activating the spray cooling solenoid valve at incorrect spatial coordinates. Summary of the Invention

[0004] The purpose of this invention is to provide an adaptive control method for microclimate in forest chicken farming based on temperature sensors, which can solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] A method for adaptive regulation of microclimate for under-forest chicken farming based on temperature sensors includes: deploying multiple temperature sensor nodes and a sealed, insulated reference cavity with a built-in standard thermometer in the under-forest farming area, wherein there is no external airflow or light interference inside the reference cavity;

[0007] Each temperature sensor node is controlled to sequentially probe into the reference cavity via an electric telescopic probe at a preset cycle. The current temperature value is read while stationary in the reference cavity and subtracted from the value of the standard thermometer to obtain the absolute drift. A two-dimensional drift matrix is ​​constructed by combining the humidity history data of the temperature sensor node's location.

[0008] During the non-calibration period, the original temperature values ​​of all the temperature sensor nodes are collected, and the original temperature values ​​are corrected using the two-dimensional drift matrix corresponding to the temperature sensor nodes. The corrected temperature values ​​are then substituted into the Kriging space interpolation algorithm to generate a three-dimensional temperature field distribution map of the forest under-forest breeding area.

[0009] Based on the coordinates of the local high-temperature zone in the three-dimensional temperature field distribution map, the spray cooling solenoid valve at the corresponding coordinate position is activated in a directional manner.

[0010] Preferably, the sealed insulated door body reference cavity includes an outer heat-insulating shell, an inner temperature-equalizing metal liner, and a sandwich vacuum isolation strip, wherein the surface of the inner temperature-equalizing metal liner is coated with a low emissivity coating.

[0011] The electric telescopic probe includes multiple sections of patterned insulating sleeves with thermal conductivity below a preset threshold. During the process of the temperature sensor node being inserted into the reference cavity, the top hatch of the sealed insulated door reference cavity is opened. When the temperature sensor node enters the geometric center of the inner uniform metal liner, the top hatch is closed. At the same time, the electromagnetic shielding layer inside the electric telescopic probe is grounded to block the electromagnetic coupling interference of radio frequency signals in the forest space to the temperature sensor node.

[0012] Preferably, the step of statically reading the current temperature value in the reference cavity specifically includes: continuously acquiring the temperature sequence of the temperature sensor node at a preset sampling frequency, calculating the temperature change slope between adjacent sampling points, and determining that the temperature sensor node has reached a thermal equilibrium state when the absolute value of the temperature change slope is lower than a preset stable threshold multiple times in a row, and taking the average value of the temperature sequence under the thermal equilibrium state as the current temperature value.

[0013] The absolute drift is obtained by calculating the difference between the current temperature value and the standard thermometer reading at the same timestamp, and by removing abnormal difference data points caused by the hysteresis effect of the standard thermometer itself.

[0014] Preferably, the humidity history data includes the average relative humidity and the percentage of liquid water condensation time in the microenvironment of the temperature sensor node in the understory breeding area;

[0015] When constructing the two-dimensional drift matrix, the average relative humidity is divided into multiple humidity intervals, and the proportion of liquid water condensation time is divided into multiple condensation intervals. A drift lookup table is constructed using the humidity intervals and the condensation intervals as two-dimensional coordinate axes. Multiple absolute drift values ​​calculated within the historical period are filled into the grid nodes corresponding to the drift lookup table. For grid nodes that have not undergone actual measurement, a reverse distance weighted interpolation method based on spatial distance weight is used to complete the numerical values, thereby generating the complete two-dimensional drift matrix.

[0016] Preferably, when correcting the original temperature value using the two-dimensional drift matrix, the current relative humidity and liquid water state of the temperature sensor node are obtained in real time, the current absolute drift amount is output by matching the two-dimensional drift matrix, and the corrected temperature value is obtained by subtracting the current absolute drift amount from the original temperature value.

[0017] When substituting the corrected temperature value into the Kriging spatial interpolation algorithm, the spatial heterogeneity caused by the canopy shading of the forest trees is taken into account. An anisotropic variogram is used instead of the isotropic variogram to calculate the spatial semivariogram between sample points. The principal axis of the anisotropic variogram is set to the prevailing wind direction of the natural wind speed under the forest.

[0018] Preferably, when determining the coordinates of the local high-temperature zone based on the three-dimensional temperature field distribution map, the spatial grids in the three-dimensional temperature field distribution map whose temperature values ​​exceed a preset high-temperature warning threshold are clustered into a continuous high-temperature region, and the geometric centroid coordinates of the continuous high-temperature region are calculated as the coordinates of the local high-temperature zone.

[0019] When the spray cooling solenoid valve at the corresponding coordinate position is activated, the spray cone angle and range parameters of the spray cooling solenoid valve are obtained. All candidate solenoid valves whose coordinates of the local high temperature zone fall within the three-dimensional space coverage range formed by the spray cone angle and range parameters are calculated. Only a specified number of solenoid valves that are closest to the geometric centroid coordinates among the candidate solenoid valves are activated.

[0020] Preferably, after the electric telescopic probe retracts the temperature sensor node and the top hatch is closed, the micro semiconductor condenser dehumidifier installed inside the reference cavity of the sealed insulated door is activated. The cold end of the micro semiconductor condenser dehumidifier is attached to the bottom of the inner layer uniform temperature metal liner, and the hot end is connected to the heat dissipation fins outside the outer layer heat insulation shell through thermal conductive silicone grease.

[0021] The humidity value inside the cavity of the inner layer uniform temperature metal liner is monitored in real time. When the humidity value of the cavity drops to a preset drying threshold, the micro semiconductor condenser dehumidifier is controlled to stop operating, so as to eliminate the thermal conduction interference of the high humidity gas under the forest when the electric telescopic probe enters and exits the reference cavity on the subsequent calibration test.

[0022] Preferably, removing abnormal difference data points caused by the hysteresis effect of the standard thermometer itself specifically includes: obtaining the historical calibration curve of the standard thermometer inside the reference cavity of the sealed insulated door body, calculating the temperature change rate envelope of the historical calibration curve near the current temperature value, determining whether the standard thermometer reading is within the temperature change rate envelope, and if it exceeds the temperature change rate envelope, it is determined as the abnormal difference data point and removed.

[0023] Before each calculation of the absolute drift, the standard thermometer reading is corrected by reading the output value of the reference voltage source integrated inside the standard thermometer and comparing the deviation between the output value of the reference voltage source and the factory-calibrated reference voltage.

[0024] Preferably, when acquiring the historical humidity data, the frequency of chicken approach and the duration of stay are collected by infrared pyroelectric sensors deployed around the temperature sensor node, and the frequency of chicken approach and the duration of stay are converted into a biological heat and moisture dissipation additional coefficient.

[0025] When constructing the drift lookup table, the biological heat and moisture dissipation coefficient is superimposed as a third dimension onto the two-dimensional coordinate axis formed by the humidity range and the condensation range to generate a three-dimensional drift lookup table. The grid node values ​​in the three-dimensional drift lookup table not only reflect the drift effect of the background humidity on the temperature sensor node, but also include the accelerated aging drift of the sensor probe surface microstructure caused by the surge in local microenvironment humidity due to chicken respiration and body surface heat dissipation.

[0026] Preferably, before calculating the candidate solenoid valves and preparing to open the specified number of solenoid valves, three-dimensional lidar point cloud data deployed in the forest breeding area is acquired, and the three-dimensional lidar point cloud data is subjected to ground filtering and clustering segmentation processing to extract the three-dimensional spatial location point set of the chicken flock in the forest.

[0027] Calculate the spatial intersection ratio between the three-dimensional spatial coverage area corresponding to the specified number of solenoid valves and the three-dimensional spatial location point set of the chicken flock. When the spatial intersection ratio exceeds a preset safety intervention threshold, adjust the installation pitch angle of the specified number of solenoid valves or reduce the number of solenoid valves opened until the spatial intersection ratio is lower than the preset safety intervention threshold, so as to avoid the spray water droplets directly contacting the feathers of the chicken flock and causing local hypothermia stress response.

[0028] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0029] 1. This invention deploys a sealed, insulated reference cavity with a built-in standard thermometer in the forest breeding area, free from external airflow and light interference. Temperature sensor nodes are controlled to sequentially probe into the reference cavity at preset intervals to read the current temperature value and compare it with the standard thermometer reading to obtain the absolute drift. A two-dimensional drift matrix is ​​constructed by combining this with historical humidity data of the location. During non-calibration periods, the original temperature values ​​are corrected using this two-dimensional drift matrix. The corrected temperature values ​​are then used in a Kriging space interpolation algorithm to generate a three-dimensional temperature field distribution map. Based on the coordinates of local high-temperature zones in this distribution map, the corresponding spray cooling solenoid valves are activated. This scheme offsets the long-term irreversible drift caused by the high humidity and corrosive environment of the forest undergrowth, corrects the spatial distribution deviation of the original temperature data, improves the accuracy of the three-dimensional temperature field reconstruction, and enables the system to activate cooling equipment based on the actual coordinates of local high-temperature zones, avoiding control misalignment caused by temperature field assessment distortion.

[0030] 2. The sealed insulated door reference cavity adopts an outer heat-insulating shell, an inner temperature-equalizing metal liner, and a vacuum isolation layer with a low-emissivity coating. Combined with the grounded electromagnetic shielding layer of the electrically telescopic probe and the operation of a micro-semiconductor condenser dehumidifier, it eliminates interference from forest undergrowth radio frequency signals and the heat conduction interference from high-humidity gas carried by the probe during the calibration process, ensuring the physical purity of the absolute drift calculation. By determining that the temperature change slope is below a preset stability threshold to confirm the thermal equilibrium state, and by eliminating abnormal data points caused by the hysteresis effect of the standard thermometer and performing secondary linear compensation correction, the accuracy of the current temperature value with the absolute drift calculation is improved. The accuracy of drift calculation was improved; the frequency and duration of chicken approach were converted into biological heat and moisture dissipation coefficients and incorporated into a three-dimensional drift lookup table; and an anisotropic variability function with the main axis direction being the dominant wind direction of natural wind speed was used during spatial interpolation to ensure that the drift matrix and temperature field reconstruction were consistent with the heterogeneity and biological disturbance characteristics of the forest understory; the spatial location point set of the chicken flock was extracted from three-dimensional lidar point cloud data, the spatial intersection ratio between the spray coverage area and the chicken flock point set was calculated, and the opening parameters of the solenoid valve were adjusted when the preset safety intervention threshold was exceeded, thus eliminating the physical risk of local hypothermia stress caused by direct contact of water droplets with the chicken flock. Attached Figure Description

[0031] Figure 1 This is an overall flowchart of a forest-based microclimate adaptive regulation method for chicken farming based on a temperature sensor, provided in an embodiment of the present invention.

[0032] Figure 2 A flowchart illustrating the thermal balance determination of a temperature sensor node entering a reference cavity, as provided in an embodiment of the present invention.

[0033] Figure 3 This is a flowchart of abnormal data removal and secondary compensation in the calculation of absolute drift provided in an embodiment of the present invention;

[0034] Figure 4 A flowchart illustrating the construction of a three-dimensional drift lookup table incorporating a biological heat and moisture dissipation coefficient, as provided in an embodiment of the present invention.

[0035] Figure 5 A flowchart for generating a three-dimensional temperature field distribution map based on anisotropic variability function provided in an embodiment of the present invention;

[0036] Figure 6 The flowchart of the control of the spray cooling solenoid valve based on point cloud data for preventing stress in chicken flocks is provided in the embodiments of the present invention. Detailed Implementation

[0037] The technical solution of the present invention will be clearly and completely described below. The described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0038] Please refer to Figure 1 This embodiment provides a sealed, insulated reference cavity containing multiple temperature sensor nodes and built-in standard thermometers deployed in a forest-based breeding area. The reference cavity is free from external airflow and light interference. Specifically, the forest-based breeding area is a large-scale free-range area with tree canopy coverage. The microenvironment within this area exhibits spatial heterogeneity based on canopy coverage, ground vegetation status, and the activity range of the chickens. The deployment locations of the temperature sensor nodes cover all zones within the forest-based breeding area that possess differentiated microenvironmental characteristics. At least one temperature sensor node is installed in each microenvironment zone to achieve full-range temperature data collection coverage of the breeding area. The reference chamber is located at the geometric center of the understory aquaculture area. The canopy coverage at this location is within the average range of the overall canopy coverage of the aquaculture area, and there are no fixed local heat sources, no continuous directional strong airflow, and no continuous direct sunlight. The reference chamber adopts a sealed and heat-insulated structure, and the interior of the chamber is thermally isolated from the external environment through the heat insulation structure. The sealed structure of the chamber prevents external airflow from entering the interior, and the outer shell of the chamber adopts an opaque structure to block external sunlight from entering the interior, so as to ensure that the temperature environment inside the chamber is not disturbed by external airflow and light. A standard thermometer is installed inside the reference chamber. The metrological performance of the standard thermometer is traceable to the national standard, so as to serve as the reference value source for temperature calibration.

[0039] Furthermore, each temperature sensor node is controlled to sequentially extend into the reference cavity via an electrically operated telescopic probe at a preset cycle. The preset cycle is set based on the environmental humidity variations in the forest aquaculture area and the temperature and humidity fluctuations caused by seasonal changes. When the relative humidity is consistently higher than the preset range, the time interval is shortened; when the relative humidity is consistently lower than the preset range, the time interval is extended to adapt to the changing drift rates of the temperature sensor nodes under different environmental conditions. Each temperature sensor node is equipped with an independent electrically operated telescopic probe. The fixed end of the probe is installed at the deployment position of the corresponding temperature sensor node, and the free end is fixedly connected to the temperature sensor node. The probe can drive the temperature sensor node to move along a preset path from its deployment position into the reference cavity. When scheduling temperature sensor nodes for calibration, each temperature sensor node is scheduled individually in order of increasing linear distance from its deployment location to the reference cavity. The next temperature sensor node's calibration scheduling process is only initiated after the previous node has completed its calibration operation and returned to its deployment location. This avoids continuous disturbances to the internal temperature environment of the reference cavity caused by simultaneous operation of multiple temperature sensor nodes, ensuring the consistency of the reference environment for each calibration operation. Figure 2 .

[0040] In this embodiment, after the temperature sensor node is inserted into the reference cavity along with the electric telescopic probe, it reads the current temperature value while stationary within the reference cavity. The current temperature value is then subtracted from the reading of a standard thermometer at the same timestamp to obtain the absolute drift of the temperature sensor node. Specifically, after the temperature sensor node is inserted into the reference cavity, the electric telescopic probe stops moving, keeping the temperature sensor node stationary at a preset position within the reference cavity, eliminating interference from airflow disturbances and frictional heat generation during movement. While stationary, the temperature data output by the temperature sensor node is continuously collected to obtain the current temperature value. Simultaneously, the reading of the standard thermometer at the same time is collected. The difference between the current temperature value and the standard thermometer reading is calculated to obtain the absolute drift of the temperature sensor node within the current calibration period. The formula for calculating the absolute drift is:

[0041]

[0042] in, Let be the absolute drift of the i-th temperature sensor node during this calibration cycle. This represents the current temperature value read by the i-th temperature sensor node when it is stationary within the reference cavity. To and The readings of the standard thermometer inside the reference cavity at the same timestamp.

[0043] A two-dimensional drift matrix is ​​constructed by combining historical humidity data from the location of the temperature sensor node. Specifically, the historical humidity data refers to the humidity-related data of the forest microenvironment collected by the temperature sensor node at its deployment location over a historical period. There is a correlation between the historical humidity data and the absolute drift of the temperature sensor node; the absolute drift of the same temperature sensor node exhibits differentiated characteristics under different humidity conditions. When constructing the two-dimensional drift matrix, two core feature dimensions of the historical humidity data are used as two-dimensional coordinate axes: one axis represents the relative humidity of the environment, and the other represents the feature dimension related to liquid water condensation in the environment. Each dimension is divided into multiple continuous intervals, forming a two-dimensional grid structure, with each grid node corresponding to a set of humidity feature parameters. The absolute drift amount obtained by calibrating the temperature sensor node under the corresponding humidity characteristic parameter combination within the historical period is filled into the corresponding grid node in the two-dimensional grid structure to form an initial two-dimensional drift data structure. For the grid nodes in the initial two-dimensional drift data structure that are not filled with the measured absolute drift amount, numerical interpolation method is used to complete the numerical data, and finally a complete two-dimensional drift matrix is ​​generated. Each element in the two-dimensional drift matrix corresponds to the absolute drift amount of the temperature sensor node under a set of humidity characteristic parameter combinations.

[0044] Table 1 Example of a two-dimensional drift matrix for a temperature sensor node

[0045] Average relative humidity range 0%-5% of condensation time 5%-10% of the condensation time 10%-15% of the condensation time 15%-20% of the condensation time 20%-25% of the condensation time 20%-30% 0.12 0.15 0.18 0.21 0.24 30%-40% 0.14 0.17 0.20 0.23 0.26 40%-50% 0.16 0.19 0.22 0.25 0.28 50%-60% 0.18 0.21 0.24 0.27 0.30 60%-70% 0.21 0.24 0.27 0.30 0.33 70%-80% 0.24 0.27 0.30 0.33 0.36 80%-90% 0.27 0.30 0.33 0.36 0.39 90%-100% 0.30 0.33 0.36 0.39 0.42

[0046] The table above shows the two-dimensional drift matrix of the i-th temperature sensor node, constructed based on the measured absolute drift and corresponding historical humidity data over six consecutive calibration cycles. The rows in the table correspond to the intervals of average relative humidity, and the columns correspond to the intervals of liquid water condensation time. The value in each cell is the absolute drift under the corresponding humidity feature combination, in °C. The values ​​of grid nodes that were not measured are completed using the inverse distance weighted interpolation method. This two-dimensional drift matrix is ​​used to correct the original temperature value of the temperature sensor node in non-calibration cycles.

[0047] During the non-calibration period, raw temperature values ​​from all temperature sensor nodes are collected, and the raw temperature values ​​are corrected using the two-dimensional drift matrix of the corresponding temperature sensor node. Specifically, the non-calibration period is the time interval between two adjacent calibration periods. During the non-calibration period, all temperature sensor nodes are in their deployment positions, continuously collecting raw temperature values ​​of their surrounding microenvironment at a preset collection frequency. The raw temperature values ​​are the temperature data directly output by the temperature sensor nodes without drift correction. Simultaneously with acquiring the raw temperature value of each temperature sensor node, the current humidity characteristic parameters of the node's location are acquired, including the current relative humidity value and the percentage of liquid water condensation time in the current period. The current humidity characteristic parameters are matched with the two-dimensional drift matrix corresponding to the temperature sensor node to obtain the current absolute drift amount under this humidity characteristic parameter combination. The raw temperature value is then corrected using the current absolute drift amount to obtain the corrected temperature value. The formula for calculating the temperature value correction is:

[0048]

[0049] in, The corrected temperature value for the i-th temperature sensor node. The original temperature value collected by the i-th temperature sensor node. This represents the current absolute drift of the i-th temperature sensor node under the current humidity characteristic parameters, obtained by matching the two-dimensional drift matrix.

[0050] The corrected temperature values ​​are then used in a Kriging spatial interpolation algorithm to generate a three-dimensional temperature field distribution map of the forest understory aquaculture area. Specifically, the three-dimensional space of the understory aquaculture area is first divided into multiple equal-volume spatial grid cells. The size of each grid cell is set based on the area of ​​the understory aquaculture area and the deployment density of temperature sensor nodes. The vertex of each grid cell serves as a spatial interpolation point for the temperature value to be estimated. The three-dimensional spatial coordinates of all temperature sensor node deployment locations are used as known sample points, and the corrected temperature values ​​of the corresponding temperature sensor nodes are used as the attribute values ​​of these known sample points. The Kriging spatial interpolation algorithm is then used to perform an unbiased optimal estimation of the temperature value at each spatial interpolation point, yielding the estimated temperature value for each point. Based on the estimated temperature values ​​at the vertices of all spatial grid cells, a three-dimensional temperature field data volume of the understory aquaculture area is constructed. A three-dimensional temperature field distribution map of the understory aquaculture area is then generated based on this data volume. This map visually displays the temperature distribution at any location within the three-dimensional space of the understory aquaculture area.

[0051] The core of the Kriging spatial interpolation algorithm is to perform linear, unbiased, optimal estimation of the attribute values ​​of the points to be estimated based on the known spatial correlation of the sample points. The formula for calculating the estimated value is as follows:

[0052]

[0053] in, The estimated temperature value for the spatial interpolation point to be estimated. The weight coefficients are the weights corresponding to the j-th known sample point. Here, represents the corrected temperature value for the j-th known sample point, and n represents the total number of known sample points, with weighting coefficients... The solution must satisfy the unbiasedness condition and the minimum variance condition. The unbiasedness condition is that the sum of all weight coefficients is 1, and the minimum variance condition is that the estimated variance of the estimated value is minimized.

[0054] The weighting coefficients are solved using a system of equations constructed from the variogram. The variogram describes the spatial correlation between known sample points, and its calculation formula is as follows:

[0055]

[0056] in, Let N(h) be the semivariance value corresponding to the lag distance h, where h is the spatial distance between two known sample points, and N(h) is the number of known sample point pairs with a lag distance of h. and These are the corrected temperature values ​​for two known sample points with a lag distance of h.

[0057] The semivariance between known sample points is calculated using the above variogram function. A set of Kriging equations is constructed, and the weight coefficients corresponding to each point to be estimated are solved. Then, the estimated temperature value of the point to be estimated is calculated. After estimating the temperature values ​​of all interpolation points in the three-dimensional space, a complete three-dimensional temperature field data volume is generated. Based on the three-dimensional temperature field data volume, a three-dimensional temperature field distribution map of the forest under-forest breeding area is generated.

[0058] Based on the coordinates of local high-temperature zones in the 3D temperature field distribution map, the corresponding spray cooling solenoid valves are activated in a targeted manner. Specifically, based on the generated 3D temperature field distribution map, local high-temperature zones are identified. These are spatial areas in the 3D temperature field where the temperature value exceeds a preset high-temperature warning threshold. Spatial clustering is performed on the identified local high-temperature zones, grouping consecutive high-temperature grid units into the same local high-temperature region. The spatial coordinates corresponding to each local high-temperature region are calculated. Multiple spray cooling solenoid valves are pre-deployed within the forest under-forest breeding area. Each spray cooling solenoid valve corresponds to a fixed 3D spatial coverage area, and their deployment positions cover the entire forest under-forest breeding area. The control terminal of each spray cooling solenoid valve is communicatively connected to the main control device, which can send start and stop control commands to the corresponding spray cooling solenoid valve. After obtaining the coordinates of the local high-temperature zones, the corresponding spray cooling solenoid valves are matched, and the valves at the corresponding coordinate positions are activated in a targeted manner to provide spray cooling intervention to the local high-temperature areas, achieving adaptive regulation of the microclimate in the forest under-forest breeding area.

[0059] In this embodiment, a unified calibration reference environment is provided for each temperature sensor node through a reference cavity, enabling in-situ periodic calibration of the temperature sensor nodes and obtaining the absolute drift of each temperature sensor node. A two-dimensional drift matrix is ​​constructed by combining historical humidity data to achieve dynamic correction of the original temperature values ​​of the temperature sensor nodes, eliminating temperature data deviation caused by sensor drift. Based on the corrected temperature values, a three-dimensional temperature field distribution map is generated using a Kriging space interpolation algorithm, ensuring the accuracy of temperature field reconstruction. Based on the accurate coordinates of local high-temperature areas, the spray cooling solenoid valve is activated, realizing precise adaptive control of the microclimate in the understory breeding area and avoiding temperature field reconstruction distortion and control misalignment caused by temperature data drift.

[0060] In one optional embodiment, the sealed insulated door reference cavity includes an outer heat-insulating shell, an inner temperature-equalizing metal liner, and a sandwich vacuum isolation strip. The surface of the inner temperature-equalizing metal liner is coated with a low emissivity coating. The electric telescopic probe includes multiple sections of patterned insulating sleeves with thermal conductivity lower than a preset threshold. During the process of the temperature sensor node being inserted into the reference cavity, the top hatch of the sealed insulated door reference cavity is opened. When the temperature sensor node enters the geometric center position of the inner temperature-equalizing metal liner, the top hatch is closed. At the same time, the electromagnetic shielding layer inside the electric telescopic probe is grounded to block the electromagnetic coupling interference of the forest space radio frequency signal to the temperature sensor node.

[0061] Specifically, the outer insulating shell is made of insulating material with low thermal conductivity, forming the external protective structure of the reference cavity and blocking the conduction of heat from the external environment into the cavity. The inner homogeneous metal liner is made of metal material with high thermal conductivity. The cavity structure of the liner is a symmetrical closed structure. The high thermal conductivity of the metal material allows the internal temperature of the liner to be quickly and evenly distributed, forming a homogeneous environment. Both the inner and outer surfaces of the liner are coated with a low emissivity coating, which reduces heat radiation exchange on the surface of the liner and reduces radiative heat transfer between the inside of the liner and the external environment, ensuring the stability of the internal temperature environment of the liner. A vacuum isolation strip is set between the outer insulating shell and the inner homogeneous metal liner. The vacuum isolation strip is a vacuum environment, which can eliminate gas heat conduction and heat convection in the interlayer, further improving the insulation performance of the reference cavity and blocking the interference of external temperature changes on the internal temperature environment of the inner homogeneous metal liner. The top of the reference cavity is equipped with an automatically opening and closing top hatch. The size of the top hatch is adapted to the outer diameter of the electric telescopic probe. When the top hatch is closed, the reference cavity is completely sealed, blocking external airflow and light from entering the cavity. When the top hatch is open, the electric telescopic probe can drive the temperature sensor node through the opening of the top hatch into the inner homogenized metal liner.

[0062] The multi-section patterned insulating sleeve of the electric telescopic probe is made of insulating material with a thermal conductivity lower than a preset threshold. This reduces heat conduction between the probe and the external environment, preventing temperature changes in the probe itself from interfering with temperature acquisition at the temperature sensor node. The patterned surface structure of the insulating sleeve increases air resistance, reducing airflow during probe extension and retraction, and minimizing airflow disturbance to the internal environment of the reference cavity. The inner wall of the electric telescopic probe is equipped with a continuous electromagnetic shielding layer made of conductive metal. After the temperature sensor node enters the geometric center of the inner homogenizing metal liner, the electromagnetic shielding layer is connected to the grounding terminal for grounding. The grounded electromagnetic shielding layer forms a continuous electromagnetic shielding space, blocking electromagnetic coupling interference from various radio frequency signals in the forest canopy to the temperature sensor node. This prevents electromagnetic interference from causing deviations in the temperature data output by the temperature sensor node, ensuring the accuracy of temperature acquisition during calibration.

[0063] In this embodiment, reading the current temperature value while stationary within the reference cavity specifically includes: continuously acquiring the temperature sequence of the temperature sensor node at a preset sampling frequency; calculating the temperature change slope between adjacent sampling points; determining that the temperature sensor node has reached thermal equilibrium when the absolute value of the temperature change slope is lower than a preset stability threshold multiple times consecutively; and taking the average of the temperature sequence under thermal equilibrium as the current temperature value. The absolute drift is obtained by calculating the difference between the current temperature value and the standard thermometer reading at the same timestamp, and removing abnormal difference data points caused by the hysteresis effect of the standard thermometer itself.

[0064] Specifically, after the temperature sensor node remains stationary at the geometric center of the inner homogeneous metal liner, it continuously collects temperature data at a preset sampling frequency, generating a continuous temperature sequence. Each data point in the temperature sequence corresponds to a unique acquisition timestamp. For the generated temperature sequence, the slope of the temperature change between two adjacent sampling points is calculated point by point. The formula for calculating the slope of the temperature change is:

[0065]

[0066] in, Let be the slope of the temperature change corresponding to the t-th sampling point. Let t be the temperature value collected at the t-th sampling point. This represents the temperature value collected at the (t-1)th sampling point. This represents the sampling time interval between two adjacent sampling points.

[0067] The slope of the temperature change sequence is continuously calculated. When the absolute value of the slope of the temperature change corresponding to multiple consecutive sampling points is lower than a preset stability threshold, it is determined that the temperature sensor node and the environment inside the inner homogeneous metal liner have reached a thermal equilibrium state. At this time, the temperature acquisition value of the temperature sensor node no longer changes significantly over time, and the acquired temperature data can accurately reflect the true temperature inside the reference cavity. After determining that a thermal equilibrium state has been reached, the continuously acquired temperature sequence under thermal equilibrium conditions is extracted, and the arithmetic mean of the temperature sequence is calculated. This arithmetic mean is used as the current temperature value of the temperature sensor node in this calibration process. By calculating the mean, random noise interference during the temperature acquisition process can be eliminated, further improving the accuracy of the current temperature value.

[0068] Table 2. Correspondence between temperature change slope and determination of thermal equilibrium state

[0069] Sampling sequence number Temperature reading (°C) Temperature change slope (°C / s) Is it below the stable threshold? Determination of thermal equilibrium state 1 24.352 - - Not achieved 2 24.387 0.035 no Not achieved 3 24.412 0.025 no Not achieved 4 24.428 0.016 no Not achieved 5 24.437 0.009 no Not achieved 6 24.442 0.005 no Not achieved 7 24.444 0.002 yes Not achieved 8 24.445 0.001 yes Not achieved 9 24.445 0.000 yes Not achieved 10 24.444 -0.001 yes Not achieved 11 24.445 0.001 yes Not achieved 12 24.445 0.000 yes Not achieved 13 24.444 -0.001 yes Not achieved 14 24.445 0.001 yes achieve 15 24.445 0.000 yes achieve 16 24.444 -0.001 yes achieve 17 24.445 0.001 yes achieve 18 24.445 0.000 yes achieve 19 24.445 0.000 yes achieve 20 24.444 -0.001 yes achieve

[0070] The table above shows the calculation results of the temperature change slope corresponding to the continuously collected temperature sequence after a temperature sensor node is stationary in the reference cavity, and the process of determining the thermal equilibrium state. The preset stability threshold is 0.002℃ / s. When the absolute value of the temperature change slope of 8 consecutive sampling points is lower than this stability threshold, it is determined that the thermal equilibrium state has been reached. The average value of the temperature sequence under the thermal equilibrium state is extracted as the current temperature value. In this embodiment, the average value of the temperature sequence under the thermal equilibrium state is 24.4447℃, that is, the current temperature value of the temperature sensor node is 24.4447℃.

[0071] After acquiring the current temperature value, the reading of a standard thermometer at the same timestamp is simultaneously acquired. The difference between the current temperature value and the standard thermometer reading is calculated to obtain the initial absolute drift data point. Due to the inherent temperature hysteresis effect of the standard thermometer, when the temperature inside the reference chamber changes slightly, the standard thermometer reading will have a delayed response, causing some difference data points to be abnormal. Therefore, it is necessary to perform outlier removal processing on the initial absolute drift data points to eliminate abnormal difference data points caused by the hysteresis effect of the standard thermometer itself.

[0072] In this embodiment, the process of eliminating abnormal difference data points caused by the hysteresis effect of the standard thermometer itself specifically includes: acquiring the historical calibration curve of the standard thermometer inside the reference cavity of the sealed insulated door body; calculating the temperature change rate envelope of the historical calibration curve near the current temperature value; determining whether the standard thermometer reading is within the temperature change rate envelope; if it exceeds the temperature change rate envelope, it is determined as an abnormal difference data point and eliminated; before each calculation of the absolute drift, by reading the output value of the reference voltage source integrated inside the standard thermometer, comparing the deviation between the reference voltage source output value and the factory-calibrated reference voltage, and performing a secondary linear compensation correction on the standard thermometer reading, referencing... Figure 3 .

[0073] Specifically, the historical calibration curve of the standard thermometer is a curve showing the change of temperature readings over time within a stable temperature environment inside the reference chamber during the historical calibration period. Based on the historical calibration curve, the maximum and minimum values ​​of the temperature change rate of the standard thermometer in different temperature ranges are calculated, forming the temperature change rate envelope for the corresponding temperature range. The temperature change rate envelope characterizes the reasonable range of the temperature change rate of the standard thermometer under normal response conditions. During this calibration process, the real-time temperature change rate of the standard thermometer reading is calculated and compared with the temperature change rate envelope corresponding to the current temperature value. If the real-time temperature change rate exceeds the upper or lower limit of the temperature change rate envelope, the standard thermometer reading at that time stamp is determined to be abnormal due to hysteresis. The corresponding absolute drift data point is an abnormal difference data point and is discarded. Only the normal data points within the temperature change rate envelope are retained for the final absolute drift calculation.

[0074] Furthermore, before each calculation of the absolute drift, a secondary linear compensation correction is performed on the standard thermometer reading to eliminate the reading deviation caused by the drift of the internal reference voltage source of the standard thermometer. Specifically, the standard thermometer integrates a reference voltage source, which provides a reference voltage for the temperature acquisition circuit. Drift in the output value of the reference voltage source directly leads to a deviation in the standard thermometer reading. During each calibration process, the real-time output value of the internal reference voltage source of the standard thermometer is read and compared with the reference voltage value calibrated at the factory. The voltage deviation between the two is calculated, and a linear compensation model is constructed based on the voltage deviation value to perform a secondary linear compensation correction on the real-time reading of the standard thermometer. The corrected standard thermometer reading can more accurately reflect the true temperature inside the reference cavity, further improving the accuracy of the absolute drift calculation. The formula for the linear compensation correction is:

[0075]

[0076] in, This is the reading after correction from a standard thermometer. This is the original reading of the standard thermometer. The voltage-to-temperature conversion factor of a standard thermometer. This is the real-time output value of the reference voltage source. This is the reference voltage value calibrated at the factory for the standard thermometer.

[0077] In this embodiment, the historical humidity data includes the average relative humidity and the proportion of liquid water condensation time in the microenvironment of the temperature sensor node in the understory breeding area. When constructing the two-dimensional drift matrix, the average relative humidity is divided into multiple humidity intervals, and the proportion of liquid water condensation time is divided into multiple condensation intervals. A drift lookup table is constructed using the humidity intervals and condensation intervals as two-dimensional coordinate axes. Multiple absolute drift values ​​calculated within the historical period are filled into the grid nodes corresponding to the drift lookup table. For grid nodes that have not undergone actual measurement, a reverse distance weighted interpolation method based on spatial distance weight is used to complete the numerical values, generating a complete two-dimensional drift matrix.

[0078] Specifically, the average relative humidity is the arithmetic mean of the relative humidity of the microenvironment where the temperature sensor node is located within a preset statistical period. The liquid water condensation time percentage is the proportion of time during which liquid water condenses on the probe surface of the temperature sensor node within the preset statistical period, relative to the total duration of the statistical period. Average relative humidity and liquid water condensation time percentage are the two core humidity characteristic dimensions affecting the drift of the temperature sensor node. When constructing the drift lookup table, the range of average relative humidity is divided into multiple continuous and non-overlapping humidity intervals. The width of each humidity interval can be set according to the variation characteristics of the average relative humidity. Similarly, the range of liquid water condensation time percentage is divided into multiple continuous and non-overlapping condensation intervals. The width of each condensation interval can be set according to the variation characteristics of the liquid water condensation time percentage. A two-dimensional grid structure drift lookup table is constructed using the humidity interval as one axis of a two-dimensional coordinate system and the condensation interval as the other axis. Each grid node in the drift lookup table corresponds to a combination of a humidity interval and a condensation interval, i.e., a set of humidity characteristic parameters.

[0079] The absolute drift obtained from the calibration of the temperature sensor node under the corresponding humidity characteristic parameter combination within the historical calibration period is filled into the corresponding grid node in the drift lookup table to complete the filling of measured data. For grid nodes in the drift lookup table that are not filled with measured absolute drift, i.e., nodes that have not undergone actual calibration measurements under the humidity characteristic parameter combination, numerical completion is performed using the inverse distance weighted interpolation method based on spatial distance weights. The core of the inverse distance weighted interpolation method is that the value of the grid node to be interpolated is obtained by weighted averaging of the values ​​of surrounding grid nodes that have been filled with measured data. The weight is inversely proportional to the spatial distance between the grid node to be interpolated and the already filled nodes; the closer the distance, the greater the weight. The calculation formula for inverse distance weighted interpolation is:

[0080]

[0081] in, Let be the absolute drift of the grid node to be interpolated. Let be the absolute drift of the m-th grid node that has been filled with measured data. Let be the weight coefficient corresponding to the m-th filled grid node, where M is the total number of filled measured data grid nodes participating in the interpolation calculation. The formula for calculating the weight coefficient is:

[0082]

[0083] in, denoted as , where is the two-dimensional spatial distance between the grid node to be interpolated and the m-th filled grid node, and p is the distance exponent, used to adjust the degree of influence of distance on the weight.

[0084] By using the aforementioned reverse distance weighted interpolation method, the values ​​of all unmeasured grid nodes in the drift lookup table are completed. The completed drift lookup table is a complete two-dimensional drift matrix. The two-dimensional drift matrix can cover all possible combinations of humidity characteristic parameters, providing complete drift data support for temperature value correction during non-calibration periods.

[0085] In this embodiment, after the electric telescopic probe retracts the temperature sensor node and the top hatch is closed, the micro-semiconductor condenser dehumidifier installed inside the reference cavity of the sealed insulated door is activated. The cold end of the micro-semiconductor condenser dehumidifier is attached to the bottom of the inner uniform temperature metal liner, and the hot end is connected to the heat dissipation fins outside the outer heat insulation shell through thermal conductive silicone grease. The humidity value of the cavity inside the inner uniform temperature metal liner is monitored in real time. When the humidity value of the cavity drops to a preset drying threshold, the micro-semiconductor condenser dehumidifier is stopped to eliminate the thermal conduction interference of the high humidity gas from the forest undergrowth brought in when the electric telescopic probe enters and exits the reference cavity on subsequent calibration tests.

[0086] Specifically, the miniature semiconductor condensing dehumidifier achieves condensation dehumidification based on the Peltier effect. Its cold end is tightly fitted to the bottom of the inner homogeneous metal liner, cooling the gas inside and causing water vapor to condense into liquid water, thus dehumidifying the cavity. The hot end is tightly connected to the heat dissipation fins on the outer insulating shell via thermally conductive silicone grease, quickly dissipating the heat generated during dehumidification to the external environment and preventing heat accumulation from interfering with the temperature environment inside the inner homogeneous metal liner. During the process of the electric telescopic probe moving the temperature sensor node in and out of the reference cavity, high-humidity gas from the forest environment is carried into the cavity. The presence of this high-humidity gas alters the heat conduction characteristics inside the cavity and forms condensation on the inner wall of the inner homogeneous metal liner and the probe surface of the temperature sensor node, interfering with subsequent calibration tests and causing deviations in the calculation of the absolute drift. Therefore, after each calibration operation is completed, once the electric telescopic probe retracts the temperature sensor node to its deployment position and the top hatch of the reference chamber is completely closed, the micro-semiconductor condenser dehumidifier is immediately activated to dehumidify the cavity inside the inner homogeneous metal liner. During dehumidification, a humidity sensor installed inside the reference chamber monitors the humidity level inside the inner homogeneous metal liner in real time. When the humidity level drops to a preset drying threshold, the micro-semiconductor condenser dehumidifier stops operating, restoring the humidity environment inside the reference chamber to a dry and stable state. This eliminates interference from entrained high-humidity gases on subsequent calibration tests and ensures the consistency of the reference environment for each subsequent calibration operation.

[0087] In this embodiment, the combination of an outer heat-insulating shell, an inner temperature-equalizing metal liner, a vacuum isolation layer, and a low-emissivity coating enhances the stability and uniformity of the internal temperature environment of the reference cavity. The low thermal conductivity insulating sleeve and grounded electromagnetic shielding layer of the electric telescopic probe eliminate the influence of heat conduction and electromagnetic interference on the calibration process. The slope of temperature change determines the thermal equilibrium state, improving the accuracy of the current temperature value acquisition. Quadratic linear compensation and abnormal data removal of the standard thermometer reading eliminate the deviation caused by the standard thermometer hysteresis effect and reference voltage drift. The inverse distance weighted interpolation method completes the two-dimensional drift matrix, ensuring its integrity. A micro-semiconductor condenser dehumidifier eliminates interference from entrained high-humidity gases, further improving the accuracy of absolute drift calculation and the stability of the calibration process, providing more accurate reference data for temperature value correction.

[0088] In another optional embodiment, when correcting the original temperature value using a two-dimensional drift matrix, the current relative humidity and liquid water state of the temperature sensor node are obtained in real time. The current absolute drift amount is matched with the two-dimensional drift matrix output. The original temperature value is subtracted from the current absolute drift amount to obtain the corrected temperature value. When the corrected temperature value is substituted into the Kriging spatial interpolation algorithm, the spatial heterogeneity characteristics caused by the canopy shading of the forest trees are taken into account. An anisotropic variogram is used instead of the isotropic variogram to calculate the spatial semivariance between sample points. The principal axis direction of the anisotropic variogram is set to the prevailing wind direction of the natural wind speed under the forest.

[0089] Specifically, during non-calibration periods, each temperature sensor node simultaneously acquires the current relative humidity of its microenvironment and the percentage of liquid water condensation time within the current statistical period, i.e., the current liquid water state, while acquiring the raw temperature value. The current relative humidity value and the percentage of liquid water condensation time are matched with the two-dimensional drift matrix corresponding to the temperature sensor node to determine the grid node corresponding to the current humidity characteristic parameter. The absolute drift amount corresponding to this grid node is extracted as the current absolute drift amount. If the current humidity characteristic parameter does not fall on a grid node of the two-dimensional drift matrix but is located in the region between four adjacent grid nodes, bilinear interpolation is used to calculate the current absolute drift amount at that location, ensuring the matching degree between the current absolute drift amount and the real-time humidity environment. The raw temperature value is corrected using the current absolute drift amount to obtain the corrected temperature value. The corrected temperature value eliminates the system deviation caused by sensor drift and can accurately reflect the true temperature of the microenvironment in which the temperature sensor node is located.

[0090] In this embodiment, the temperature spatial correlation in the understory aquaculture area varies significantly in different directions due to the shading effect of the tree canopy. Furthermore, the prevailing wind direction of the natural wind speed under the trees causes the temperature field to exhibit different spatial variation characteristics along the wind direction and perpendicular to the wind direction; that is, the spatial heterogeneity of the temperature field exhibits anisotropic characteristics. Traditional isotropic variograms assume that the spatial correlation of the temperature field is consistent in all directions, which cannot adapt to the anisotropic characteristics of the understory aquaculture area and will lead to biases in the Kriging interpolation results. Therefore, anisotropic variograms are used instead of isotropic variograms to calculate the spatial semivariance between known sample points.

[0091] Specifically, when calculating the semivariogram, the anisotropic variogram considers not only the spatial distance between two sample points but also their spatial direction. Different variogram parameters are set for different directions to accommodate variations in the spatial correlation of temperature. The principal axis of the anisotropic variogram is set to the prevailing annual wind direction of the natural wind speed under the forest canopy, representing the direction with the strongest spatial correlation in the forest temperature field and the largest range of the variogram. The secondary axis, perpendicular to the principal axis, represents the direction with the weakest spatial correlation and the smallest range of the variogram. The formula for calculating the anisotropic variogram is:

[0092]

[0093] in, Here, θ represents the semivariance value corresponding to the lag distance h and direction θ, where θ is the angle between the line connecting the two sample points and the principal axis direction. C is the nugget value, and C is the sill value. For the theoretical model of the variogram, Let θ be the effective lag distance. The formula for calculating the effective lag distance is:

[0094]

[0095] in, The range of motion in the main axis direction. denoted as the range along the secondary axis, and h is the Euclidean distance between the two sample points.

[0096] By using the anisotropic variogram, the semivariogram values ​​of known sample points in different directions are calculated. A Kriging equation system adapted to the anisotropic characteristics of the forest canopy is constructed, and the weighting coefficients corresponding to each spatial interpolation point to be estimated are obtained. The estimated temperature value of the point is then calculated. The Kriging spatial interpolation algorithm using the anisotropic variogram can fully adapt to the spatial heterogeneity of the temperature field caused by the canopy shading of trees and the prevailing wind direction, improving the accuracy of the three-dimensional temperature field reconstruction. This allows the generated three-dimensional temperature field distribution map to more realistically reflect the temperature distribution of the forest canopy aquaculture area. (Reference) Figure 5 .

[0097] In this embodiment, when determining the coordinates of a local high-temperature zone based on the three-dimensional temperature field distribution map, the spatial grids in the three-dimensional temperature field distribution map whose temperature values ​​exceed the preset high-temperature warning threshold are clustered into a continuous high-temperature region, and the geometric centroid coordinates of the continuous high-temperature region are calculated as the coordinates of the local high-temperature zone. When the spray cooling solenoid valve at the corresponding coordinate position is activated in a directional manner, the spray cone angle and range parameters of the spray cooling solenoid valve are obtained, and all candidate solenoid valves whose coordinates of the local high-temperature zone fall within the three-dimensional spatial coverage range formed by the spray cone angle and range parameters are calculated. Only a specified number of solenoid valves closest to the geometric centroid coordinates among the candidate solenoid valves are activated.

[0098] Specifically, the preset high-temperature warning threshold is an upper limit of temperature set based on the suitable growth temperature range for chickens raised under forest cover. In the three-dimensional temperature field data volume corresponding to the three-dimensional temperature field distribution map, each spatial grid cell corresponds to a unique temperature value. All spatial grid cells in the three-dimensional temperature field data volume are traversed, and high-temperature grid cells with temperature values ​​exceeding the preset high-temperature warning threshold are selected. Spatial clustering is performed on the selected high-temperature grid cells, grouping spatially adjacent and connected high-temperature grid cells into the same continuous high-temperature region. Each continuous high-temperature region corresponds to an independent local high-temperature zone. For each clustered continuous high-temperature region, the arithmetic mean of the three-dimensional spatial coordinates of all high-temperature grid cells within that region is calculated to obtain the geometric centroid coordinates of that continuous high-temperature region. These geometric centroid coordinates are used as the local high-temperature zone coordinates, accurately representing the core location of the local high-temperature zone.

[0099] Each spray cooling solenoid valve deployed within the forest aquaculture area corresponds to a fixed spray cone angle and range parameter. The spray cone angle is the size of the cone-shaped spray area formed by the spray water flow after the solenoid valve is opened. The range parameter is the maximum spray distance of the spray water flow under a preset pressure. Based on the spray cone angle and range parameter, the three-dimensional spatial coverage area corresponding to each spray cooling solenoid valve can be determined. This three-dimensional spatial coverage area is a conical spatial region with the installation position of the solenoid valve as the vertex, the spray cone angle as the cone angle, and the range parameter as the height. After obtaining the geometric centroid coordinates of the local high-temperature zone, the three-dimensional spatial coverage areas of all spray cooling solenoid valves are traversed, and all spray cooling solenoid valves whose geometric centroid coordinates fall within their three-dimensional spatial coverage area are selected as candidate solenoid valves. Among the candidate solenoid valves, they are sorted in order of increasing spatial distance between the installation position of the solenoid valve and the geometric centroid coordinates of the local high-temperature zone. A specified number of solenoid valves at the top of the sorted list are selected, and only the selected specified number of solenoid valves are opened to perform directional spray cooling on the local high-temperature zone. This method allows for the activation of only the minimum number of solenoid valves covering the core area of ​​the localized high-temperature zone, avoiding water waste and excessive humidity increases in the overall aquaculture area caused by widespread valve activation. Simultaneously, it enables precise cooling intervention in the localized high-temperature zone. (Refer to...) Figure 6 .

[0100] Table 3. Matching Relationship between the Three-Dimensional Spatial Coverage Range of the Spray Cooling Solenoid Valve and the Local High Temperature Zone

[0101] Solenoid valve number 3D coordinates of installation location (m) Spray cone angle (°) Range parameters (m) Are the geometric barycentric coordinates within the coverage area? Is it a candidate solenoid valve? Distance from the geometric centroid (m) Enable or disable? 01 (12.5,8.2,2.1) 60 3.5 no no 4.2 no 02 (15.3,10.1,2.0) 60 3.5 yes yes 1.8 yes 03 (17.6,9.4,2.1) 60 3.5 yes yes 2.5 no 04 (14.8,12.7,2.0) 60 3.5 yes yes 3.1 no 05 (18.2,13.1,2.1) 60 3.5 no no 5.3 no 06 (11.7,11.2,2.0) 60 3.5 no no 3.7 no 07 (16.4,7.8,2.1) 60 3.5 no no 3.9 no 08 (19.1,11.5,2.0) 60 3.5 no no 4.6 no

[0102] The table above shows the matching results between the geometric centroid coordinates of a local high-temperature area and the corresponding spray cooling solenoid valve. The geometric centroid coordinates of this local high-temperature area are (15.8, 10.7, 0.2). The preset high-temperature warning threshold is 30℃, and the specified number of valves to be opened is 1. By matching the geometric centroid coordinates with the three-dimensional spatial coverage of the solenoid valves, three candidate solenoid valves are selected. Finally, the solenoid valve No. 02, which is closest to the geometric centroid coordinates, is opened to achieve targeted and precise cooling of the local high-temperature area.

[0103] In this embodiment, when acquiring historical humidity data, infrared pyroelectric sensors deployed around the temperature sensor node collect the frequency and duration of chicken approach and stay, converting these parameters into a bio-thermal and moisture dissipation coefficient. When constructing the drift lookup table, the bio-thermal and moisture dissipation coefficient is superimposed as a third dimension onto the two-dimensional coordinate axis formed by the humidity and condensation intervals, generating a three-dimensional drift lookup table. The grid node values ​​in the three-dimensional drift lookup table not only reflect the background humidity's impact on the temperature sensor node's drift but also include the accelerated aging drift caused by the surge in local microenvironment humidity due to chicken respiration and surface heat dissipation, affecting the microstructure of the sensor probe surface. Figure 4 .

[0104] Specifically, infrared pyroelectric sensors are deployed within a preset range around each temperature sensor node. They can detect flocks of chickens entering this range, collect the number of times the flock enters the range (i.e., the frequency of the flock approaching), and the duration of the flock's stay within the range each time it enters. The flock's respiration and heat dissipation from its body surface cause a sudden surge in the local microenvironment humidity around the temperature sensor node. At the same time, the excrement and other waste generated by the flock's activities exacerbate the corrosivity of the surrounding environment, accelerate the aging of the microstructure on the surface of the temperature sensor node probe, and cause the sensor's drift rate to increase and the drift amount to increase.

[0105] The frequency and duration of chicken approach data were converted into a bio-thermal and moisture emission coefficient. This coefficient is positively correlated with both the frequency and duration of chicken approach; higher frequency and longer duration result in a larger coefficient, indicating a greater influence of chicken activity on sensor drift. In constructing the drift lookup table, the bio-thermal and moisture emission coefficient was added as a third dimension to the existing two dimensions: average relative humidity and the proportion of liquid water condensation time. A three-dimensional coordinate axis was constructed, dividing the coefficient's value range into multiple continuous intervals, forming a three-dimensional grid structure. Each grid node in the lookup table corresponds to a set of parameters: average relative humidity, the proportion of liquid water condensation time, and the bio-thermal and moisture emission coefficient.

[0106] The absolute drift value obtained from the calibration of the temperature sensor node under the corresponding parameter combination during the historical calibration period is filled into the corresponding grid node in the three-dimensional drift lookup table. For grid nodes without measured data, a three-dimensional backward distance weighted interpolation method is used to complete the values, generating a complete three-dimensional drift lookup table. The value of each grid node in the three-dimensional drift lookup table not only includes the influence of the ambient background humidity on the drift of the temperature sensor node, but also the accelerating influence of local microenvironmental changes caused by chicken activity on sensor drift. This allows for a more comprehensive adaptation to the sensor drift characteristic changes brought about by chicken activity in the forest-based farming environment. During non-calibration periods, the current absolute drift value can be obtained by combining real-time collected chicken activity data with the three-dimensional drift lookup table, further improving the accuracy of temperature value correction.

[0107] In this embodiment, before calculating the candidate solenoid valves and preparing to open a specified number of solenoid valves, three-dimensional lidar point cloud data deployed in the forest breeding area is acquired. The three-dimensional lidar point cloud data is subjected to ground filtering and clustering segmentation to extract the three-dimensional spatial location point set of the chicken flock. The spatial intersection ratio between the three-dimensional spatial coverage area corresponding to the specified number of solenoid valves and the three-dimensional spatial location point set of the chicken flock is calculated. When the spatial intersection ratio exceeds a preset safety intervention threshold, the installation pitch angle of the specified number of solenoid valves is adjusted or the number of solenoid valves opened is reduced until the spatial intersection ratio is lower than the preset safety intervention threshold, so as to avoid the spray droplets directly contacting the chicken flock's feathers and causing local hypothermia stress response.

[0108] Specifically, multiple 3D LiDARs are deployed within the forest-based breeding area. The scanning range of these LiDARs covers the entire breeding area, enabling real-time acquisition of 3D LiDAR point cloud data. This point cloud data contains the 3D spatial coordinates of all objects within the breeding area. After acquiring candidate solenoid valves and determining the specified number of valves to be opened, the 3D point cloud data of the forest-based breeding area collected by the LiDARs at the current moment is acquired. This point cloud data undergoes preprocessing. First, ground filtering is performed to separate ground points from non-ground points, removing interference from ground points. Then, non-ground points are clustered. Based on the spatial distance and geometric features of the point cloud, point clouds belonging to the same object are clustered into a single point cloud set. Based on the 3D geometric features of the chicken flock, point clouds belonging to the chicken flock are extracted from the clustered point cloud set, forming a 3D spatial location point set for the chicken flock. This point set accurately represents the 3D spatial distribution of the chicken flock within the forest-based breeding area at the current moment.

[0109] For a specified number of solenoid valves to be activated, the three-dimensional spatial coverage area corresponding to each solenoid valve is determined. The three-dimensional spatial coverage areas of all solenoid valves to be activated are then merged to form the total spray coverage area. The spatial intersection ratio between the total spray coverage area and the set of three-dimensional spatial locations of the chickens is calculated. The spatial intersection ratio is the proportion of the number of points in the chickens' three-dimensional spatial location set that fall within the total spray coverage area to the total number of points in the chickens' three-dimensional spatial location set. This ratio characterizes the degree of overlap between the spray coverage area and the chickens' locations; a higher ratio indicates a greater probability of direct contact between the spray droplets and the chickens. The formula for calculating the spatial intersection ratio is:

[0110]

[0111] in, For the spatial intersection ratio, This refers to the number of points within the total spray coverage area where the three-dimensional spatial location points of the chicken flock converge. This represents the total number of points in the three-dimensional spatial location set of the chicken flock.

[0112] The preset safety intervention threshold is a proportional upper limit set based on the stress response characteristics of the chicken flock. When the calculated spatial intersection ratio exceeds the preset safety intervention threshold, it indicates that the spray coverage area of ​​the solenoid valve to be opened overlaps too much with the location of the chicken flock. The probability of the spray droplets directly contacting the chicken feathers is high, which will trigger a local hypothermia stress response in the chicken flock. Therefore, it is necessary to adjust the opening parameters of the solenoid valve. There are two adjustment methods. One is to adjust the installation pitch angle of the solenoid valve to be opened. The installation pitch angle of the solenoid valve is the angle between the spray direction of the solenoid valve and the horizontal direction. By adjusting the installation pitch angle, the three-dimensional spatial coverage area of ​​the solenoid valve can be changed, so that the total spray coverage area avoids the three-dimensional spatial location point set of the chicken flock, reducing the spatial intersection ratio. The other is to reduce the number of solenoid valves to be opened, retaining only the solenoid valves that are closest to the geometric centroid coordinates of the local high temperature area and whose spray coverage area does not overlap with the location of the chicken flock or has a very low degree of overlap, thereby reducing the degree of overlap between the total spray coverage area and the location of the chicken flock. The opening parameters of the solenoid valves are continuously adjusted until the spatial intersection ratio between the adjusted total spray coverage area and the three-dimensional spatial location set of the chickens is lower than a preset safety intervention threshold. Then, a start control command is sent to the adjusted solenoid valves to open them. In this way, precise cooling of local high-temperature areas can be achieved while avoiding direct contact between spray droplets and the chickens' feathers. This eliminates the risk of localized hypothermia stress caused by spray contact and ensures the healthy growth of chickens raised in the forest.

[0113] In this embodiment, the Kriging space interpolation algorithm based on anisotropic variograms fully adapts to the spatial heterogeneity of the temperature field in the understory breeding area, improving the accuracy of the three-dimensional temperature field reconstruction. By calculating the geometric centroid coordinates of local high-temperature zones and matching and screening candidate solenoid valves, the directional and precise activation of the spray cooling solenoid valve is achieved. By adding a three-dimensional drift lookup table with the additional coefficient of biological heat and moisture dissipation of the chicken flock, the accuracy of temperature sensor node drift correction is further improved. By extracting the spatial position of the chicken flock from three-dimensional lidar point cloud data, calculating the spatial intersection ratio between the spray coverage area and the chicken flock position, and adjusting the solenoid valve opening parameters, the hypothermia stress response caused by direct contact of spray droplets with the chicken flock is avoided, further improving the accuracy and safety of the adaptive regulation of the microclimate in understory breeding.

Claims

1. A method for adaptive regulation of microclimate in forest-raised chickens based on temperature sensors, characterized in that, include: Multiple temperature sensor nodes and a sealed, insulated reference cavity with a built-in standard thermometer are deployed in the forest breeding area. The reference cavity is free from external airflow and light interference. Each temperature sensor node is controlled to sequentially probe into the reference cavity via an electric telescopic probe at a preset cycle. The current temperature value is read while stationary in the reference cavity and subtracted from the value of the standard thermometer to obtain the absolute drift. A two-dimensional drift matrix is ​​constructed by combining the humidity history data of the temperature sensor node's location. During the non-calibration period, the original temperature values ​​of all the temperature sensor nodes are collected, and the original temperature values ​​are corrected using the two-dimensional drift matrix corresponding to the temperature sensor nodes. The corrected temperature values ​​are then substituted into the Kriging space interpolation algorithm to generate a three-dimensional temperature field distribution map of the forest under-forest breeding area. Based on the coordinates of the local high-temperature zone in the three-dimensional temperature field distribution map, the spray cooling solenoid valve at the corresponding coordinate position is activated in a directional manner.

2. The adaptive regulation method for forest-based chicken farming microclimate based on temperature sensors according to claim 1, characterized in that, The sealed insulated door body reference cavity includes an outer heat-insulating shell, an inner temperature-equalizing metal liner, and a sandwich vacuum isolation strip. The surface of the inner temperature-equalizing metal liner is coated with a low emissivity coating. The electric telescopic probe includes multiple sections of patterned insulating sleeves with thermal conductivity below a preset threshold. During the process of the temperature sensor node being inserted into the reference cavity, the top hatch of the reference cavity of the sealed insulated door is opened. When the temperature sensor node enters the geometric center of the inner uniform temperature metal liner, the top hatch is closed, and at the same time, the electromagnetic shielding layer inside the electric telescopic probe is grounded.

3. The adaptive regulation method for forest-based chicken farming microclimate based on temperature sensors according to claim 1, characterized in that, The step of statically reading the current temperature value in the reference cavity specifically includes: continuously acquiring the temperature sequence of the temperature sensor node at a preset sampling frequency, calculating the temperature change slope between adjacent sampling points, and determining that the temperature sensor node has reached a thermal equilibrium state when the absolute value of the temperature change slope is lower than a preset stable threshold multiple times in a row, and taking the average value of the temperature sequence under the thermal equilibrium state as the current temperature value. The absolute drift is obtained by calculating the difference between the current temperature value and the standard thermometer reading at the same timestamp, and by removing abnormal difference data points caused by the hysteresis effect of the standard thermometer itself.

4. The adaptive regulation method for forest-based chicken farming microclimate based on temperature sensors according to claim 1, characterized in that, The humidity history data includes the average relative humidity and the percentage of liquid water condensation time in the microenvironment of the temperature sensor node in the understory breeding area. When constructing the two-dimensional drift matrix, the average relative humidity is divided into multiple humidity intervals, and the proportion of liquid water condensation time is divided into multiple condensation intervals. A drift lookup table is constructed using the humidity intervals and the condensation intervals as two-dimensional coordinate axes. Multiple absolute drift values ​​calculated within the historical period are filled into the grid nodes corresponding to the drift lookup table. For grid nodes that have not undergone actual measurement, a reverse distance weighted interpolation method based on spatial distance weight is used to complete the numerical values, thereby generating the complete two-dimensional drift matrix.

5. The adaptive regulation method for forest-based chicken farming microclimate based on temperature sensors according to claim 1, characterized in that, When correcting the original temperature value using the two-dimensional drift matrix, the current relative humidity and liquid water state of the temperature sensor node are obtained in real time, the current absolute drift amount is output by matching the two-dimensional drift matrix, and the corrected temperature value is obtained by subtracting the current absolute drift amount from the original temperature value. When substituting the corrected temperature value into the Kriging spatial interpolation algorithm, the spatial heterogeneity caused by the canopy shading of the forest trees is taken into account. An anisotropic variogram is used instead of the isotropic variogram to calculate the spatial semivariogram between sample points. The principal axis of the anisotropic variogram is set to the prevailing wind direction of the natural wind speed under the forest.

6. The adaptive regulation method for forest-based chicken farming microclimate based on temperature sensors according to claim 1, characterized in that, When determining the coordinates of the local high-temperature zone based on the three-dimensional temperature field distribution map, the spatial grids in the three-dimensional temperature field distribution map whose temperature values ​​exceed the preset high-temperature warning threshold are clustered into a continuous high-temperature region, and the geometric centroid coordinates of the continuous high-temperature region are calculated as the coordinates of the local high-temperature zone. When the spray cooling solenoid valve at the corresponding coordinate position is activated, the spray cone angle and range parameters of the spray cooling solenoid valve are obtained. All candidate solenoid valves whose coordinates of the local high temperature zone fall within the three-dimensional space coverage range formed by the spray cone angle and range parameters are calculated. Only a specified number of solenoid valves that are closest to the geometric centroid coordinates among the candidate solenoid valves are activated.

7. The temperature sensor based microclimate self-adaptive regulation method for chicken farming under forest according to claim 2, characterized in that, After the electric telescopic probe retracts the temperature sensor node and the top hatch is closed, the micro semiconductor condenser dehumidifier installed inside the reference cavity of the sealed insulated door is activated. The cold end of the micro semiconductor condenser dehumidifier is attached to the bottom of the inner layer uniform temperature metal liner, and the hot end is connected to the heat dissipation fins outside the outer layer heat insulation shell through thermal conductive silicone grease. The humidity value inside the cavity of the inner layer uniform temperature metal liner is monitored in real time. When the humidity value of the cavity drops to a preset drying threshold, the micro semiconductor condenser dehumidifier is controlled to stop operating.

8. The adaptive regulation method for forest-based chicken farming microclimate based on temperature sensors according to claim 3, characterized in that, The process of removing abnormal difference data points caused by the hysteresis effect of the standard thermometer itself specifically includes: obtaining the historical calibration curve of the standard thermometer inside the reference cavity of the sealed insulated door body, calculating the temperature change rate envelope of the historical calibration curve near the current temperature value, determining whether the reading of the standard thermometer is within the temperature change rate envelope, and if it exceeds the temperature change rate envelope, it is determined as the abnormal difference data point and removed. Before each calculation of the absolute drift, the standard thermometer reading is corrected by reading the output value of the reference voltage source integrated inside the standard thermometer and comparing the deviation between the output value of the reference voltage source and the factory-calibrated reference voltage.

9. The temperature sensor based microclimate self-adaptive regulation method for chicken farming under forest according to claim 4, characterized in that, When acquiring the historical humidity data, the frequency of chicken approach and the duration of stay are collected by infrared pyroelectric sensors deployed around the temperature sensor node, and the frequency of chicken approach and the duration of stay are converted into a biological heat and moisture dissipation additional coefficient. When constructing the drift lookup table, the biological heat and moisture dissipation coefficient is superimposed as a third dimension onto the two-dimensional coordinate axis formed by the humidity range and the condensation range to generate a three-dimensional drift lookup table. The grid node values ​​in the three-dimensional drift lookup table not only reflect the drift effect of the background humidity on the temperature sensor node, but also include the accelerated aging drift of the sensor probe surface microstructure caused by the surge in local microenvironment humidity due to chicken respiration and body surface heat dissipation.

10. The adaptive regulation method for forest-based chicken farming microclimate based on temperature sensors according to claim 6, characterized in that, Before calculating the candidate solenoid valves and preparing to open the specified number of solenoid valves, three-dimensional lidar point cloud data deployed in the forest breeding area is obtained, and the three-dimensional lidar point cloud data is subjected to ground filtering and clustering segmentation to extract the three-dimensional spatial location point set of the chicken flock in the forest. Calculate the spatial intersection ratio between the three-dimensional spatial coverage area corresponding to the specified number of solenoid valves and the three-dimensional spatial location point set of the chicken flock. When the spatial intersection ratio exceeds a preset safety intervention threshold, adjust the installation pitch angle of the specified number of solenoid valves or reduce the number of solenoid valves opened until the spatial intersection ratio is lower than the preset safety intervention threshold, so as to avoid the spray water droplets directly contacting the feathers of the chicken flock and causing local hypothermia stress response.

Images