A method and system for controlling the temperature in a greenhouse

Abstract

The application discloses a greenhouse temperature control method and system, and the method comprises the following steps: acquiring real-time temperature data of different crop growth areas, a global temperature field distribution image of a greenhouse, and external environment data of the greenhouse; performing data fusion processing on the real-time temperature data and the temperature field distribution image, establishing a global temperature matrix, and combining the external environment data of the greenhouse to construct a temperature influence factor model; predicting temperature variation trends of each area of the greenhouse in a future preset time period, comparing current real-time temperatures of each area with corresponding current suitable temperature thresholds of planted crops, determining temperature deviation areas, and calculating temperature regulation differences of each deviation area; and regulating the temperature of each deviation area according to the spatial position of each area, the type of crops and the temperature regulation difference, until the temperature of the deviation area of the greenhouse crops is stably within the corresponding suitable temperature threshold range. The application solves the problem of low regulation accuracy of the greenhouse temperature control method in the prior art.

Description

Technical Field

[0001] This invention relates to the field of greenhouse temperature control technology, and in particular to a greenhouse temperature control method and system. Background Technology

[0002] With the rapid development of modern agricultural technology, greenhouse cultivation, as a highly efficient agricultural production model, has been widely applied to the large-scale cultivation of various crops such as vegetables, fruits, and flowers due to its advantages of artificially intervening in the growth environment and overcoming the limitations of natural climate. Greenhouse cultivation not only effectively extends the crop growth cycle and improves yield and quality per unit area, but also plays an irreplaceable role in ensuring the supply of agricultural products during the off-season, coping with extreme weather disasters, and promoting the development of agricultural industrialization. Among the various environmental control factors in greenhouse cultivation, temperature is the core factor directly determining the crop's growth status, physiological metabolism, and final output. Each growth stage of a crop, such as seed germination, stem and leaf growth, flowering, and fruiting, corresponds to a strict suitable temperature threshold range. When the temperature inside the greenhouse is below the lower threshold, the crop's cellular respiration weakens and enzyme activity decreases, leading to growth stagnation, frost damage, or even death; when the temperature is above the upper threshold, transpiration intensifies, nutrient consumption is too rapid, and problems such as excessive vegetative growth, flower and fruit drop, and high incidence of pests and diseases easily occur. Therefore, achieving precise and stable temperature control inside the greenhouse is a key link in ensuring healthy crop growth and improving planting efficiency.

[0003] To meet the temperature control requirements of greenhouses, traditional technologies have developed various control methods. In the early days, farmers mainly relied on manual adjustment based on their planting experience, using equipment such as roller shutters, fans, and heaters to regulate the temperature. This method not only consumes a lot of labor costs, but also heavily depends on personal experience for control accuracy. It is difficult to cope with dynamic temperature changes and is prone to excessive temperature fluctuations due to delayed judgment or operational errors, making it impossible to provide a stable growing environment for crops.

[0004] With the initial application of sensing technology, some greenhouses have begun to use single-point or a small number of distributed temperature sensors for temperature monitoring, combined with simple temperature control logic to achieve semi-automatic regulation. For example, when the sensor detects that the temperature is higher than the set value, the fan is automatically started to ventilate and cool down; when the temperature is lower than the set value, the heating equipment is triggered to operate.

[0005] However, such technologies have significant limitations: Firstly, the monitoring range of a single point or a small number of sensors is limited, failing to comprehensively cover different crop growth areas within the greenhouse. This is especially problematic for larger greenhouses, where temperature monitoring blind spots easily appear, making it difficult to reflect the overall temperature distribution differences. Consequently, "one-size-fits-all" control measures cannot adapt to the individualized needs of crops in different areas. Secondly, relying solely on single temperature data within the greenhouse as the basis for control fails to consider the influence of external environmental factors (such as outside temperature, light intensity, and wind speed) on the temperature inside the greenhouse. This makes it impossible to predict temperature changes in advance, and control strategies often lag behind actual temperature fluctuations, easily leading to energy waste or untimely control. Therefore, current greenhouse temperature control methods still suffer from low accuracy. Summary of the Invention

[0006] In view of this, the purpose of the present invention is to provide a greenhouse temperature control method and system, which aims to solve the problem that the existing greenhouse temperature control methods still have low control accuracy.

[0007] The embodiments of the present invention are implemented as follows: One aspect proposes a method for controlling greenhouse temperature, the method comprising: Real-time temperature data of different crop growth areas collected by distributed temperature sensors inside the greenhouse, temperature field distribution images of the entire greenhouse collected by infrared thermal imaging devices, and environmental data outside the greenhouse collected by environmental monitoring modules are acquired respectively. Data fusion processing is performed on real-time temperature data and temperature field distribution images. A global temperature matrix is ​​established by mapping the pixel coordinates to the sensor position. A temperature influencing factor model is constructed by combining the external environmental data of the greenhouse. Based on the temperature influence factor model, predict the temperature change trend of each area of ​​the greenhouse within a preset time period, compare the current real-time temperature of each area with the corresponding suitable temperature threshold for the current crop, determine the temperature deviation area, and calculate the temperature control difference of each deviation area. Based on the spatial location, crop type, and temperature control difference of each region, the temperature of each deviation region is adjusted until the temperature of the deviation region where greenhouse crops are growing stabilizes within the corresponding suitable temperature threshold range.

[0008] Furthermore, in the aforementioned greenhouse temperature control method, the step of performing data fusion processing on real-time temperature data and temperature field distribution images, and establishing a global temperature matrix through the mapping relationship between pixel coordinates and sensor positions, includes: Adaptive filtering and noise reduction are performed on the real-time temperature data collected by distributed temperature sensors. The temperature fluctuation characteristics reflecting the crop growth microenvironment in the sensor data are preserved through a dual verification mechanism of sliding window and temperature gradient threshold. The temperature field distribution image acquired by the infrared thermal imaging device is subjected to grayscale stretching and edge enhancement. The effective pixel area corresponding to the crop growth area in the temperature field image is extracted by the crop canopy contour recognition algorithm, and invalid temperature pixels are removed. Using the pre-set three-dimensional coordinate system inside the greenhouse as a reference, the spatial coordinate parameters of each distributed temperature sensor are obtained. Through the internal parameter calibration and external parameter calibration of the infrared thermal imaging device, the conversion relationship between pixel coordinates in the temperature field distribution image and the three-dimensional spatial coordinates of the greenhouse is established. The RANSAC random sampling consensus algorithm is adopted, and the transformation relationship between pixel coordinates and spatial coordinates is optimized with the sensor spatial coordinates and their corresponding temperature data as constraints. A precise mapping model between sensor position and image pixels is constructed to realize the association assignment of sensor temperature data to the corresponding pixel region. Based on the optimized pixel-sensor mapping model, the sensor temperature data is filled into the effective pixel area corresponding to the temperature field image to form an initial temperature matrix. For pixel areas in the initial temperature matrix that are not directly covered by the sensor, an interpolation algorithm is used to calculate and fill in the temperature data of the blank pixels by combining the temperature values ​​of adjacent valid pixels, crop type difference coefficients and spatial distance weights. Finally, the corresponding global temperature matrix is ​​generated through matrix dimension standardization.

[0009] Furthermore, in the aforementioned greenhouse temperature control method, the step of employing the RANSAC random sampling consensus algorithm, using sensor spatial coordinates and their corresponding temperature data as constraints, to optimize the transformation relationship between pixel coordinates and spatial coordinates, constructing a precise mapping model between sensor positions and image pixels, and realizing the associative assignment of sensor temperature data to corresponding pixel regions includes: Clean the sensor spatial coordinates and corresponding temperature data, and bind the sensor spatial coordinates with temperature values ​​to form constraint data in a preset format to establish a constraint dataset. Randomly select some samples from the constrained dataset, substitute them into the pixel coordinate and spatial coordinate transformation relationship to calculate the corresponding pixel coordinates, compare the calculated pixel coordinates with the actual pixel coordinates of the corresponding sensor area in the image, and obtain the projection error. Set an error threshold, classify samples with projection errors less than the threshold as inliers, and the rest as outliers. Iterate and repeat the process of extracting some samples, calculating errors, and filtering inliers. In each round, use the current set of inliers to reconstruct the pixel coordinates and spatial coordinates until the proportion of inliers is stable above the preset ratio. Output the final pixel coordinates and spatial coordinates transformation relationship. By combining the sensor coordinates with the final pixel coordinates and spatial coordinates transformation relationship, a bidirectional mapping channel is formed to construct a precise mapping model between the sensor position and image pixels. The bidirectional mapping channel includes a forward channel and a reverse channel. The forward channel outputs the corresponding image pixel coordinates by inputting the sensor's three-dimensional coordinates, while the reverse channel infers the corresponding three-dimensional spatial coordinates of the greenhouse by inputting the image pixel coordinates.

[0010] Furthermore, in the aforementioned greenhouse temperature control method, the steps of using an interpolation algorithm to calculate and complete the temperature data of blank pixels by combining the temperature values ​​of adjacent valid pixels, crop type difference coefficients, and spatial distance weights for pixel areas not directly covered by sensors in the initial temperature matrix, and finally generating the corresponding global temperature matrix through matrix dimension standardization, include: Collect temperature sensitivity thresholds and temperature change tolerance data for various crop types at different growth stages in the greenhouse, and establish a mapping relationship table of crop type-growth stage-temperature response characteristics; Based on the actual growth stage of the crop corresponding to each blank pixel area, the matching temperature response coefficient is extracted from the mapping table, and the initially set crop type difference coefficient is dynamically corrected to obtain the dynamic difference coefficient that is updated in real time with the growth stage. Centered on the blank pixel, a neighborhood range with a radius of R is defined, and the temperature data of all valid pixels within the neighborhood range are extracted; A spatial distance attenuation factor is introduced, and the dynamic difference coefficient and the spatial distance attenuation factor are weighted and fused to obtain a comprehensive weight. The temperature value of each effective pixel is fitted and calculated with the corresponding comprehensive weight, and the initial predicted temperature of the blank pixel is output and filled into the initial temperature matrix. Using the three-dimensional coordinate system of the greenhouse X axis, Y Using axes as the basis for the row and column dimensions of the matrix, and unifying the pixel resolution of the matrix, the initial temperature matrix after completion is dimensionally normalized to generate a global temperature matrix.

[0011] Furthermore, in the above-mentioned greenhouse temperature control method, the correction formula for the dynamic difference coefficient is: , The dynamic difference coefficient. This represents the initial crop type difference coefficient. This represents the temperature-sensitive threshold for the current growth stage. The average temperature sensitivity threshold for the entire crop cycle; The formula for calculating the spatial distance attenuation factor is: , in, It is the linear distance between valid pixels and blank pixels.

[0012] Furthermore, in the above-mentioned greenhouse temperature control method, the step of constructing a temperature influencing factor model based on external greenhouse environmental data includes: Key factors in the greenhouse external environment data were screened and divided into dynamic factors and static factors. Dynamic factors include real-time wind speed, solar radiation intensity, and air humidity, while static factors include geographical location latitude and seasonal attributes. The mean value and mutation characteristics of dynamic factors are extracted over a unit time, and the static factors are converted into feature values ​​through standardization to construct an environmental feature set. Using environmental feature set as input and actual temperature change as output, a temperature influence factor model is obtained through regression algorithm training.

[0013] Furthermore, in the aforementioned greenhouse temperature control method, the step of adjusting the temperature of each deviating area based on the spatial location, crop type, and temperature regulation difference of each area until the temperature of the deviating area where the greenhouse crop grows stabilizes within the corresponding suitable temperature threshold range includes: Based on the spatial location of each area in the greenhouse, the effective control area of ​​each temperature control device is marked to form a table of correspondence between devices and areas. At the same time, the crop types in each area are recorded, and the temperature sensitivity characteristics and suitable temperature requirements of different crops are queried and determined. Based on the magnitude of the temperature regulation difference, the deviation areas are divided into three categories: small difference, medium difference, and large difference. The regulation intensity is adjusted in combination with the crop type. Among them, the regulation is strengthened in the large difference area for warm-loving crops, and a mild regulation method is adopted in the small difference area for heat-resistant crops. Based on the equipment and area correspondence table, assign dedicated or nearby temperature control equipment to each off-center area until the temperature in the off-center area where greenhouse crops are growing stabilizes within the corresponding suitable temperature threshold range.

[0014] Another object of the present invention is to provide a greenhouse temperature control system, the system comprising: The acquisition module is used to acquire real-time temperature data of different crop growth areas collected by distributed temperature sensors inside the greenhouse, temperature field distribution images of the entire greenhouse collected by infrared thermal imaging devices, and environmental data outside the greenhouse collected by the environmental monitoring module. The fusion module is used to perform data fusion processing on real-time temperature data and temperature field distribution images. It establishes a global temperature matrix by mapping the pixel coordinates to the sensor position and constructs a temperature influence factor model by combining the external environmental data of the greenhouse. The calculation module is used to predict the temperature change trend of each area of ​​the greenhouse within a preset time period based on the temperature influence factor model, compare the current real-time temperature of each area with the corresponding suitable temperature threshold for the current crop, determine the temperature deviation area, and calculate the temperature control difference of each deviation area. The control module is used to regulate the temperature of each deviation area based on the spatial location, crop type, and temperature regulation difference of each area, until the temperature of the deviation area where the greenhouse crops are growing stabilizes within the corresponding suitable temperature threshold range.

[0015] In another aspect, the present invention provides a readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.

[0016] In another aspect, the present invention provides a computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the method described above.

[0017] This invention acquires real-time temperature data from distributed temperature sensors in different crop growth areas within a greenhouse, a greenhouse-wide temperature field distribution image from an infrared thermal imaging device, and external environmental data from an environmental monitoring module. The real-time temperature data and temperature field distribution image are fused together to establish a global temperature matrix based on the mapping relationship between pixel coordinates and sensor positions. A temperature influence factor model is constructed by combining this model with the external environmental data. Based on this model, the temperature change trend of each area within the greenhouse over a preset time period is predicted. The current real-time temperature of each area is compared with the corresponding suitable temperature threshold for the currently grown crop to identify temperature deviation areas, and the temperature control difference for each deviation area is calculated. The temperature of each deviation area is then controlled according to its spatial location, crop type, and temperature control difference until the temperature in the deviation area stabilizes within the corresponding suitable temperature threshold range. This approach no longer relies on traditional manual experience or single sensor data but combines the regional accuracy of distributed sensing with the global coverage of infrared thermal imaging to characterize the global temperature distribution. External environmental parameters serve as the core input to the temperature influence factor model, predicting temperature change trends and taking into account the influence of the external environment on the greenhouse temperature. This solves the problem that existing greenhouse temperature control methods still suffer from low accuracy. Attached Figure Description

[0018] Figure 1 This is a flowchart of the greenhouse temperature control method proposed in the first embodiment of the present invention; Figure 2 This is a schematic diagram of the greenhouse temperature control system in the third embodiment of the present invention.

[0019] The following detailed description, in conjunction with the accompanying drawings, will further illustrate the present invention. Detailed Implementation

[0020] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. Several embodiments of the invention are illustrated in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.

[0021] It should be noted that when a component is said to be "fixed to" another component, it can be directly on the other component or there may be an intervening component. When a component is said to be "connected to" another component, it can be directly connected to the other component or there may be an intervening component. The terms "vertical," "horizontal," "left," "right," and similar expressions used in this document are for illustrative purposes only.

[0022] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0023] Example 1 Please see Figure 1 The figure shows a greenhouse temperature control method proposed in the first embodiment of the present invention, the method including steps S10 to S13.

[0024] Step S10: Acquire real-time temperature data of different crop growth areas collected by distributed temperature sensors inside the greenhouse, temperature field distribution images of the entire greenhouse collected by infrared thermal imaging devices, and environmental data outside the greenhouse collected by the environmental monitoring module.

[0025] This involves comprehensive multi-source data collection, covering key temperature and environmental data both inside and outside the greenhouse. Specifically, distributed temperature sensors deployed inside the greenhouse acquire real-time temperature data for different crop growth areas. This distributed deployment ensures data collection covers the microenvironments of different crops, avoiding the limitations of single sensors. Simultaneously, infrared thermal imaging devices capture temperature field distribution images of the entire greenhouse area. These images visually represent the spatial distribution characteristics of temperature within the greenhouse, compensating for the spatial coverage deficiencies of single-point sensors. Furthermore, an environmental monitoring module collects environmental data from outside the greenhouse. The external environment is a crucial factor influencing temperature changes inside the greenhouse; incorporating this data improves the accuracy of subsequent temperature control.

[0026] Step S11: Perform data fusion processing on real-time temperature data and temperature field distribution images, establish a global temperature matrix through the mapping relationship between pixel coordinates and sensor positions, and construct a temperature influence factor model by combining the greenhouse external environment data.

[0027] Next, the collected internal and external data are fused and a model is built. Real-time temperature data is fused with temperature field distribution images. The core of this fusion is establishing a mapping relationship between pixel coordinates and sensor positions. This mapping associates discrete sensor temperature data with continuous image pixels, forming a global temperature matrix and transforming single-point data into a global temperature distribution. Based on this, combined with previously collected greenhouse external environmental data, a temperature influencing factor model is constructed. This model integrates internal and external factors, providing a mathematical basis for subsequent temperature change prediction.

[0028] For example, when constructing a temperature influencing factor model, the first step is to screen and classify key factors from the external environmental data of the greenhouse. From the external environmental data collected by the environmental monitoring module, key factors that significantly affect temperature changes inside the greenhouse are selected and divided into two categories: dynamic factors and static factors. Dynamic factors refer to real-time changing environmental parameters, including real-time wind speed (affecting the rate of heat exchange between the inside and outside of the greenhouse), solar radiation intensity (directly affecting heat acquisition inside the greenhouse), and air humidity (affecting evaporative cooling efficiency). Static factors refer to relatively stable environmental parameters, including geographical latitude (affecting the angle and duration of solar radiation) and seasonal attributes (reflecting the overall climate characteristics of the period). This classification facilitates subsequent targeted feature extraction.

[0029] A standardized environmental feature set is constructed. For dynamic factors, their mean value per unit time (reflecting the average state over a period of time) and abrupt change characteristics (reflecting drastic changes in a short period of time, such as sudden wind or sudden clearing) are extracted. These features can comprehensively characterize the changing patterns of dynamic factors. For static factors, they are transformed into feature values ​​suitable for model input through standardization (converting parameters of different dimensions and ranges into feature values ​​of a uniform scale). The processed dynamic factor features are integrated with the static factor features to construct a complete environmental feature set, which serves as the input data for the temperature influence factor model.

[0030] A temperature influence factor model is generated by training a regression algorithm. The constructed environmental feature set is used as the input variable of the model, and the actual temperature change data occurring inside the greenhouse is used as the output variable. A suitable regression algorithm (such as linear regression, nonlinear regression, machine learning regression model, etc.) is employed to train the model on the data. Through continuous training, the model parameters are optimized so that the model can accurately capture the correlation between environmental features and temperature changes inside the greenhouse, ultimately resulting in a temperature influence factor model that can predict temperature changes based on environmental features.

[0031] Step S12: Based on the temperature influence factor model, predict the temperature change trend of each area of ​​the greenhouse within a preset time period in the future, compare the current real-time temperature of each area with the corresponding suitable temperature threshold for the current crop, determine the temperature deviation area, and calculate the temperature control difference of each deviation area.

[0032] Among these methods, temperature prediction and control parameter determination are carried out based on the constructed temperature influencing factor model. The model is used to predict the temperature change trend of each area of ​​the greenhouse within a preset time period. At the same time, the real-time temperature of each area is compared with the suitable temperature threshold of the corresponding crop to accurately locate the areas where the temperature deviates from the suitable range. For these deviating areas, the temperature difference that needs to be adjusted is further calculated, that is, the specific temperature adjustment value required to reach the suitable temperature, providing a quantitative indicator for control operations.

[0033] Specifically, the core of this step is to locate the temperature problem by combining the current state and future trends. Compared with judging solely based on real-time temperature, this approach is more forward-looking and accurate. First, the basis for determining the suitable temperature threshold is clarified. This involves combining crop type, growth stage, and planting practice data to determine a specific standard. By consulting crop cultivation data, agricultural experimental data, and local planting experience, corresponding temperature ranges are set for each crop at different stages such as germination, growth, and flowering and fruiting. This forms a "crop-growth stage-suitable temperature threshold" correlation table, which serves as a benchmark for judging whether the temperature deviates from the target range.

[0034] Secondly, the temperature influence factor model is used to output the temperature change trend within a preset time period in the future. This trend includes not only the direction of temperature change in each region (increasing, decreasing or stabilizing), but also the magnitude of change (e.g., increasing by 1°C every 30 minutes or decreasing by 0.5°C every hour).

[0035] Finally, the real-time temperature of each region is directly compared with the suitable temperature threshold for the corresponding crop to screen out "current deviation areas" where the real-time temperature is higher than the upper limit of the threshold or lower than the lower limit. Trend prediction is introduced: for regions where the current temperature is within the threshold range, the future deviation is predicted based on the temperature change trend—for example, if a region's current temperature is 25℃ (within the suitable threshold of 20-28℃ for tomato growth), but the model predicts that the temperature in that region will rise to 30℃ in the next hour due to increased solar radiation, it is preemptively classified as a "potential deviation area." For regions already in the current deviation area, the degree of deviation is judged based on the trend to see if it will worsen (e.g., if the current temperature is 18℃, lower than the lower limit of the suitable threshold for tomatoes, and the predicted temperature will continue to drop to 15℃, it is judged as a "severely deviating and continuously deteriorating area"). Through this method of "real-time temperature initial screening + trend prediction completion," existing temperature problems can be accurately located, and impending temperature risks can be identified in advance, avoiding misjudgments (such as ignoring areas that will exceed the standard in the short term) or omissions (such as mistakenly classifying temporarily normal but soon-to-deviate areas as normal) caused solely by real-time temperature, thus achieving comprehensive and accurate positioning of temperature deviation areas.

[0036] Step S13: Based on the spatial location, crop type, and temperature control difference of each area, adjust the temperature of each deviation area until the temperature of the deviation area for greenhouse crop growth stabilizes within the corresponding suitable temperature threshold range.

[0037] This involves implementing targeted temperature control operations. Personalized control strategies are developed and executed based on the spatial location of each deviation area, the type of crop planted, and the calculated temperature control difference. Temperature changes are continuously monitored during the control process until the temperature in all deviation areas stabilizes within the suitable temperature threshold range for their respective crops, completing a full temperature control closed loop.

[0038] For example, when implementing temperature control, the first step is to establish a correspondence between equipment and regions, as well as a table of crop temperature requirements. Based on the spatial location of each area within the greenhouse, the effective control range of each temperature control device (such as heaters, fans, and shade curtains) is determined (i.e., the area of ​​the greenhouse that the device can affect), thus forming a correspondence table between temperature control devices and control areas. Simultaneously, the types of crops planted in each area are recorded. By consulting relevant crop characteristic data or previously accumulated data, the temperature sensitivity characteristics of different crops (such as their sensitivity to low or high temperatures) and specific suitable temperature requirements are determined, providing a basis for formulating control strategies.

[0039] The deviation areas were categorized, and the intensity of regulation was determined. Based on the previously calculated temperature regulation difference, all temperature deviation areas were divided into three categories: small difference (small gap from the suitable temperature), medium difference, and large difference (large gap from the suitable temperature). The regulation intensity was adjusted according to the crop type in each region, and a differentiated regulation strategy was formulated: for large difference areas of warm-loving crops (which have high temperature requirements and whose growth is greatly affected after deviating from the suitable temperature), the regulation intensity was strengthened first to ensure a rapid recovery to the suitable temperature; for small difference areas of heat-tolerant crops (which have high tolerance to temperature fluctuations), a mild regulation method was adopted to avoid excessive regulation that would lead to energy waste or sudden temperature changes that would affect crop growth.

[0040] Assign temperature control equipment and implement regulation until the temperature stabilizes. Based on the previously established equipment-region correspondence table, assign a dedicated temperature control device (if the region has a corresponding dedicated device) or a nearby temperature control device to each off-center region (to improve regulation efficiency and reduce heat transfer loss); start the temperature control equipment to perform regulation operations, and continuously monitor temperature changes in the off-center regions. Adjust the regulation parameters in a timely manner based on temperature feedback until the temperature in all off-center regions stabilizes within the suitable temperature threshold range for the corresponding crops, thus completing the regulation process.

[0041] In summary, the greenhouse temperature control method in the above embodiments of the present invention acquires real-time temperature data of different crop growth areas collected by distributed temperature sensors inside the greenhouse, temperature field distribution images of the entire greenhouse collected by an infrared thermal imaging device, and environmental data of the outside environment collected by an environmental monitoring module. The real-time temperature data and temperature field distribution images are fused together, and a global temperature matrix is ​​established through the mapping relationship between pixel coordinates and sensor positions. A temperature influence factor model is constructed by combining the external environmental data. Based on the temperature influence factor model, the temperature change trend of each area of ​​the greenhouse within a preset time period is predicted. The current real-time temperature of each area is compared with the corresponding suitable temperature threshold for the currently planted crop to determine the temperature deviation areas, and the temperature regulation difference of each deviation area is calculated. The temperature of each deviation area is regulated according to its spatial location, crop type, and temperature regulation difference until the temperature of the deviation area for greenhouse crop growth stabilizes within the corresponding suitable temperature threshold range. No longer relying on traditional manual experience or single sensor data, this method combines the regional precision of distributed sensing with the full-area coverage of infrared thermal imaging to characterize the overall temperature distribution. External environmental parameters serve as the core input to the temperature influence factor model, predicting temperature change trends and taking into account the impact of the external environment on the internal temperature. This solves the problem of low accuracy in existing greenhouse temperature control methods.

[0042] Example 2 This embodiment also proposes a greenhouse temperature control method. The difference between the greenhouse temperature control method proposed in this embodiment and the greenhouse temperature control method proposed in Embodiment 1 is as follows: The step of performing data fusion processing on real-time temperature data and temperature field distribution images, and establishing a global temperature matrix through the mapping relationship between pixel coordinates and sensor positions, includes: Adaptive filtering and noise reduction are performed on the real-time temperature data collected by distributed temperature sensors. The temperature fluctuation characteristics reflecting the crop growth microenvironment in the sensor data are preserved through a dual verification mechanism of sliding window and temperature gradient threshold. The temperature field distribution image acquired by the infrared thermal imaging device is subjected to grayscale stretching and edge enhancement. The effective pixel area corresponding to the crop growth area in the temperature field image is extracted by the crop canopy contour recognition algorithm, and invalid temperature pixels are removed. Using the pre-set three-dimensional coordinate system inside the greenhouse as a reference, the spatial coordinate parameters of each distributed temperature sensor are obtained. Through the internal parameter calibration and external parameter calibration of the infrared thermal imaging device, the conversion relationship between pixel coordinates in the temperature field distribution image and the three-dimensional spatial coordinates of the greenhouse is established. The RANSAC random sampling consensus algorithm is adopted, and the transformation relationship between pixel coordinates and spatial coordinates is optimized with the sensor spatial coordinates and their corresponding temperature data as constraints. A precise mapping model between sensor position and image pixels is constructed to realize the association assignment of sensor temperature data to the corresponding pixel region. Based on the optimized pixel-sensor mapping model, the sensor temperature data is filled into the effective pixel area corresponding to the temperature field image to form an initial temperature matrix. For pixel areas in the initial temperature matrix that are not directly covered by the sensor, an interpolation algorithm is used to calculate and fill in the temperature data of the blank pixels by combining the temperature values ​​of adjacent valid pixels, crop type difference coefficients and spatial distance weights. Finally, the corresponding global temperature matrix is ​​generated through matrix dimension standardization.

[0043] First, the real-time temperature data collected by distributed temperature sensors is denoised using an adaptive filtering method, while simultaneously incorporating a dual verification mechanism of a sliding window and a temperature gradient threshold. The sliding window smooths the data locally, reducing the impact of instantaneous fluctuations, while the temperature gradient threshold distinguishes between normal temperature changes and abnormal noise signals. This dual verification preserves the temperature fluctuation characteristics of the sensor data that truly reflect the crop growth microenvironment, eliminates invalid noise, and ensures the reliability of the basic data.

[0044] For temperature field distribution images acquired by infrared thermal imaging devices, preprocessing is performed to extract effective information. First, grayscale stretching is used to enhance image contrast, then edge enhancement is performed to highlight contour features in the image. Subsequently, a crop canopy contour recognition algorithm is used to accurately locate the effective pixel regions in the temperature field image corresponding to the crop growth area. At the same time, invalid temperature pixels in non-crop areas (such as aisles, equipment areas, etc.) inside the greenhouse are removed to ensure that subsequent temperature data processing focuses on the core area.

[0045] A coordinate transformation relationship is established to link the image with the actual space. Using a pre-defined three-dimensional coordinate system within the greenhouse as a reference, the precise spatial coordinate parameters of each distributed temperature sensor are obtained. Through intrinsic parameter calibration (determining the device's own optical parameters) and extrinsic parameter calibration (determining the device's installation position and orientation within the greenhouse space) of the infrared thermal imaging device, a transformation relationship is established between the pixel coordinates in the temperature field distribution image and the three-dimensional spatial coordinates of the greenhouse, laying the foundation for subsequent data association.

[0046] The RANSAC (Random Sample Consensus) algorithm is used to optimize the coordinate transformation relationship and construct a mapping model. Using the sensor's spatial coordinates and their corresponding temperature data as constraints, the previously established pixel coordinate and spatial coordinate transformation relationship is optimized. This algorithm constructs a precise mapping model between the sensor position and image pixels, ultimately achieving accurate association and assignment of the temperature data collected by the sensor to the corresponding pixel region, establishing a one-to-one correspondence between temperature data and spatial pixels.

[0047] An initial temperature matrix is ​​constructed based on the optimized mapping model. The denoised sensor temperature data is then filled into the effective pixel areas corresponding to the temperature field image according to the mapping model. The resulting matrix contains only the temperature data of the sensor coverage area, forming the initial temperature matrix.

[0048] The blank areas in the initial temperature matrix are filled in and standardized. For pixel areas in the initial temperature matrix not directly covered by the sensor, an interpolation algorithm is used to fill in the temperature data. The interpolation calculation not only refers to the temperature values ​​of adjacent valid pixels, but also combines the crop type difference coefficient (different crops respond differently to temperature) and spatial distance weight (the closer the pixels are, the greater their influence) to ensure the accuracy of the filled data. After filling, the matrix is ​​standardized in terms of dimensions to unify the matrix specifications and resolution, and finally a global temperature matrix that can comprehensively and accurately reflect the temperature distribution inside the greenhouse is generated.

[0049] Furthermore, the step of employing the RANSAC random sampling consensus algorithm, using the sensor's spatial coordinates and their corresponding temperature data as constraints, to optimize the transformation relationship between pixel coordinates and spatial coordinates, constructing a precise mapping model between sensor location and image pixels, and realizing the associative assignment of sensor temperature data to corresponding pixel regions includes: Clean the sensor spatial coordinates and corresponding temperature data, and bind the sensor spatial coordinates with temperature values ​​to form constraint data in a preset format to establish a constraint dataset. Randomly select some samples from the constrained dataset, substitute them into the pixel coordinate and spatial coordinate transformation relationship to calculate the corresponding pixel coordinates, compare the calculated pixel coordinates with the actual pixel coordinates of the corresponding sensor area in the image, and obtain the projection error. Set an error threshold, classify samples with projection errors less than the threshold as inliers, and the rest as outliers. Iterate and repeat the process of extracting some samples, calculating errors, and filtering inliers. In each round, use the current set of inliers to reconstruct the pixel coordinates and spatial coordinates until the proportion of inliers is stable above the preset ratio. Output the final pixel coordinates and spatial coordinates transformation relationship. By combining the sensor coordinates with the final pixel coordinates and spatial coordinates transformation relationship, a bidirectional mapping channel is formed to construct a precise mapping model between the sensor position and image pixels. The bidirectional mapping channel includes a forward channel and a reverse channel. The forward channel outputs the corresponding image pixel coordinates by inputting the sensor's three-dimensional coordinates, while the reverse channel infers the corresponding three-dimensional spatial coordinates of the greenhouse by inputting the image pixel coordinates.

[0050] The first step is to establish a constraint dataset. The spatial coordinates of the sensors and their corresponding temperature data are cleaned, and outlier data (such as abrupt changes or missing values ​​caused by sensor malfunctions) is removed to ensure data quality. Simultaneously, the spatial coordinates of each sensor are bound to their corresponding temperature values ​​to form constraint data in a pre-defined format (such as a unified data table structure). This integrated constraint dataset provides a reliable input basis for subsequent algorithm optimization.

[0051] Then, sample extraction and projection error calculation are performed. A portion of sample data is randomly extracted from the constrained dataset, and the sensor spatial coordinates in the samples are substituted into the previously established pixel coordinates and spatial coordinate transformation relationship to calculate the corresponding pixel coordinates. Subsequently, the calculated pixel coordinates are compared with the pixel coordinates of the actual area corresponding to the sensor in the infrared thermal imaging image. The deviation between the two is the projection error, which reflects the accuracy of the current coordinate transformation relationship.

[0052] The final coordinate transformation relationship is determined through iterative optimization. A pre-set error threshold is used to classify samples with projection errors less than the threshold as inliers (valid samples that conform to the current transformation relationship), and the remaining samples as outliers (abnormal or non-conforming samples). The process of sample extraction, error calculation, and inlier / outlier selection is repeated. In each iteration, the pixel coordinate and spatial coordinate transformation relationship is recalculated and updated using the current inlier set. The iteration continues until the proportion of inliers in the total samples stabilizes above the preset proportion (indicating that the transformation relationship is accurate enough). The iteration is then stopped and the final pixel coordinate and spatial coordinate transformation relationship is output.

[0053] A precise mapping model between sensor location and image pixels is constructed. The spatial coordinates of the sensor are combined with the final determined pixel-to-spatial coordinate transformation relationship to form a bidirectional mapping channel, thus constructing a complete mapping model. The forward channel takes the sensor's three-dimensional spatial coordinates as input and outputs its corresponding pixel coordinates in the temperature field distribution image. The reverse channel, on the other hand, takes a pixel coordinate from the input image and deduces the corresponding spatial coordinates of that pixel in the greenhouse's three-dimensional coordinate system. This bidirectional mapping ensures a precise correspondence between spatial location and image pixels.

[0054] Furthermore, the steps for calculating and filling in the temperature data of blank pixels in the initial temperature matrix by using an interpolation algorithm, combining the temperature values ​​of adjacent valid pixels, crop type difference coefficients, and spatial distance weights, and finally generating the corresponding global temperature matrix through matrix dimension standardization include: Collect temperature sensitivity thresholds and temperature change tolerance data for various crop types at different growth stages in the greenhouse, and establish a mapping relationship table of crop type-growth stage-temperature response characteristics; Based on the actual growth stage of the crop corresponding to each blank pixel area, the matching temperature response coefficient is extracted from the mapping table, and the initially set crop type difference coefficient is dynamically corrected to obtain the dynamic difference coefficient that is updated in real time with the growth stage. Centered on the blank pixel, a neighborhood range with a radius of R is defined, and the temperature data of all valid pixels within the neighborhood range are extracted; A spatial distance attenuation factor is introduced, and the dynamic difference coefficient and the spatial distance attenuation factor are weighted and fused to obtain a comprehensive weight. The temperature value of each effective pixel is fitted and calculated with the corresponding comprehensive weight, and the initial predicted temperature of the blank pixel is output and filled into the initial temperature matrix. Using the X and Y axes of the greenhouse's three-dimensional coordinate system as the matrix row and column dimension references, and unifying the pixel resolution of the matrix, the initial temperature matrix after completion is dimensionally normalized to generate a global temperature matrix.

[0055] Among these methods, the accuracy of temperature data completion in blank areas is improved by incorporating crop growth characteristics and spatial distance factors.

[0056] First, a mapping table of crop temperature response characteristics was established. Key temperature parameters of various crops at different growth stages (such as germination, growth, and fruiting stages) were extensively collected in greenhouses, including temperature sensitivity thresholds (critical temperatures at which crops respond significantly to temperature changes) and temperature tolerance (the range of temperature fluctuations that crops can withstand). Based on these data, a mapping table of "crop type - growth stage - temperature response characteristics" was established, which provides a basis for subsequent temperature calculations.

[0057] Dynamically correct crop type difference coefficients. Based on the actual growth stage of the crop corresponding to each blank pixel area, extract the matching temperature response coefficient from the above mapping table; use this coefficient to dynamically correct the initially set crop type difference coefficient, obtaining a dynamic difference coefficient that can be updated in real time with the crop growth stage, ensuring that the difference coefficient matches the current growth status of the crop and improving the pertinence of temperature calculation.

[0058] Determine the effective pixel range of the neighborhood of the blank pixel. Centered on the blank pixel that needs to be filled with temperature data, define a neighborhood range with a radius of R (the value of R can be set according to the actual situation such as greenhouse size and sensor density); extract the temperature information of all effective pixels that have been filled with sensor temperature data within this neighborhood range. The temperature data of these adjacent effective pixels are the core reference for calculating the temperature of the blank pixel.

[0059] In addition, in some optional embodiments of the present invention, geometric topology can be incorporated into the blank pixel neighborhood partitioning process, abandoning the traditional fixed radius R neighborhood mode and adopting a geometric partitioning method of "Voronoi diagram (Thysen polygon) + crop planting boundary constraint"—using all distributed temperature sensors in the greenhouse as generators, and combining geometric features such as crop row boundaries and planting area outlines, a Voronoi diagram of the entire greenhouse is constructed, so that each blank pixel is uniquely assigned to the polygon area corresponding to the nearest sensor, while ensuring that the neighborhood range does not cross the planting boundaries of different crop types. This geometric partitioning can accurately match the spatial coverage of the sensors and the crop distribution characteristics, avoiding temperature data interference caused by crossing crop areas in the traditional fixed radius neighborhood.

[0060] Calculate the temperature of the blank pixels and fill it into the matrix. Introduce a spatial distance attenuation factor (the farther away an effective pixel is from the blank pixel, the smaller its temperature impact). Weight the previously obtained dynamic difference coefficient with the spatial distance attenuation factor to obtain a comprehensive weight. Fit the temperature value of each effective pixel in the neighborhood to the corresponding comprehensive weight (e.g., weighted average) to obtain the initial predicted temperature of the blank pixel, and fill the corresponding position in the initial temperature matrix with this temperature value.

[0061] The matrix is ​​dimensionally normalized to generate a global temperature matrix. The X-axis (horizontal) and Y-axis (vertical) of the greenhouse's three-dimensional coordinate system are used as the row and column dimensions of the temperature matrix, respectively, to unify the pixel resolution of the matrix (ensuring that each pixel corresponds to a fixed actual spatial area). The initial temperature matrix after filling in all blank pixel temperature data is dimensionally normalized to meet the unified specification requirements, and finally a complete and standardized global temperature matrix is ​​generated.

[0062] For example, the correction formula for the dynamic difference coefficient is: , The dynamic difference coefficient. This represents the initial crop type difference coefficient. This represents the temperature-sensitive threshold for the current growth stage. The average temperature sensitivity threshold for the entire crop cycle; The formula for calculating the spatial distance attenuation factor is: , in, It is the linear distance between valid pixels and blank pixels.

[0063] In summary, the greenhouse temperature control method proposed in the above embodiments of the present invention acquires real-time temperature data of different crop growth areas collected by distributed temperature sensors inside the greenhouse, temperature field distribution images of the entire greenhouse collected by an infrared thermal imaging device, and environmental data of the outside environment collected by an environmental monitoring module. The real-time temperature data and temperature field distribution images are fused together, and a global temperature matrix is ​​established through the mapping relationship between pixel coordinates and sensor positions. A temperature influence factor model is constructed by combining the external environmental data. Based on the temperature influence factor model, the temperature change trend of each area of ​​the greenhouse within a preset time period is predicted. The current real-time temperature of each area is compared with the corresponding suitable temperature threshold for the currently planted crop to determine the temperature deviation areas, and the temperature regulation difference of each deviation area is calculated. The temperature of each deviation area is regulated according to its spatial location, crop type, and temperature regulation difference until the temperature of the deviation area for greenhouse crop growth stabilizes within the corresponding suitable temperature threshold range. No longer relying on traditional manual experience or single sensor data, this method combines the regional precision of distributed sensing with the full-area coverage of infrared thermal imaging to characterize the overall temperature distribution. External environmental parameters serve as the core input to the temperature influence factor model, predicting temperature change trends and taking into account the impact of the external environment on the internal temperature. This solves the problem of low accuracy in existing greenhouse temperature control methods.

[0064] Example 3 Please see Figure 2 The figure shows a greenhouse temperature control system proposed in the third embodiment of the present invention, the system comprising: The acquisition module 100 is used to acquire real-time temperature data of different crop growth areas collected by distributed temperature sensors inside the greenhouse, temperature field distribution images of the entire greenhouse collected by infrared thermal imaging devices, and environmental data outside the greenhouse collected by the environmental monitoring module. The fusion module 200 is used to perform data fusion processing on real-time temperature data and temperature field distribution images. It establishes a global temperature matrix by mapping the pixel coordinates to the sensor position and constructs a temperature influence factor model by combining the greenhouse external environment data. The calculation module 300 is used to predict the temperature change trend of each area of ​​the greenhouse within a preset time period based on the temperature influence factor model, compare the current real-time temperature of each area with the corresponding suitable temperature threshold for the current crop, determine the temperature deviation area, and calculate the temperature control difference of each deviation area. The control module 400 is used to regulate the temperature of each deviation area according to the spatial location, crop type and temperature regulation difference of each area, until the temperature of the deviation area where the greenhouse crop grows is stable within the corresponding suitable temperature threshold range.

[0065] The functions or operation steps implemented by the above modules are largely the same as those in the above method embodiments, and will not be repeated here.

[0066] Example 4 In another aspect, the present invention provides a readable storage medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the steps of the method described in any one of the above embodiments one to two.

[0067] Example 5 In another aspect, the present invention provides a computer device, the computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of any one of the methods described in embodiments one to two above.

[0068] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0069] Those skilled in the art will understand that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequential list of executable instructions for implementing logical functions, and can be embodied in any computer-readable storage medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable storage medium" can mean any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0070] More specific examples (a non-exhaustive list) of computer-readable storage media include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable storage media can even be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0071] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0072] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0073] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A method of controlling the temperature in a greenhouse, characterized in that The method includes: Real-time temperature data of different crop growth areas collected by distributed temperature sensors inside the greenhouse, temperature field distribution images of the entire greenhouse collected by infrared thermal imaging devices, and environmental data outside the greenhouse collected by environmental monitoring modules are acquired respectively. Data fusion processing is performed on real-time temperature data and temperature field distribution images. A global temperature matrix is ​​established by mapping the pixel coordinates to the sensor position. A temperature influencing factor model is constructed by combining the external environmental data of the greenhouse. Based on the temperature influence factor model, predict the temperature change trend of each area of ​​the greenhouse within a preset time period, compare the current real-time temperature of each area with the corresponding suitable temperature threshold for the current crop, determine the temperature deviation area, and calculate the temperature control difference of each deviation area. Based on the spatial location, crop type, and temperature control difference of each region, the temperature of each deviation region is controlled until the temperature of the deviation region where the greenhouse crop grows stabilizes within the corresponding suitable temperature threshold range. The step of performing data fusion processing on real-time temperature data and temperature field distribution images, and establishing a global temperature matrix through the mapping relationship between pixel coordinates and sensor positions, includes: Adaptive filtering and noise reduction are performed on the real-time temperature data collected by distributed temperature sensors. The temperature fluctuation characteristics reflecting the crop growth microenvironment in the sensor data are preserved through a dual verification mechanism of sliding window and temperature gradient threshold. The temperature field distribution image acquired by the infrared thermal imaging device is subjected to grayscale stretching and edge enhancement. The effective pixel area corresponding to the crop growth area in the temperature field image is extracted by the crop canopy contour recognition algorithm, and invalid temperature pixels are removed. Using the pre-set three-dimensional coordinate system inside the greenhouse as a reference, the spatial coordinate parameters of each distributed temperature sensor are obtained. Through the internal parameter calibration and external parameter calibration of the infrared thermal imaging device, the conversion relationship between pixel coordinates in the temperature field distribution image and the three-dimensional spatial coordinates of the greenhouse is established. The RANSAC random sampling consensus algorithm is adopted, and the transformation relationship between pixel coordinates and spatial coordinates is optimized with the sensor spatial coordinates and their corresponding temperature data as constraints. A precise mapping model between sensor position and image pixels is constructed to realize the association assignment of sensor temperature data to the corresponding pixel region. Based on the optimized pixel-sensor mapping model, the sensor temperature data is filled into the effective pixel area corresponding to the temperature field image to form an initial temperature matrix. For pixel regions in the initial temperature matrix not directly covered by the sensor, an interpolation algorithm is used to calculate and fill in the temperature data of the blank pixels by combining the temperature values ​​of adjacent valid pixels, crop type difference coefficients, and spatial distance weights. Finally, through matrix dimension normalization, the corresponding global temperature matrix is ​​generated, including: Collect temperature sensitivity thresholds and temperature change tolerance data for various crop types at different growth stages in the greenhouse, and establish a mapping relationship table of crop type-growth stage-temperature response characteristics; Based on the actual crop growth stage corresponding to each blank pixel area, the matching temperature response coefficient is extracted from the mapping table, and the initially set crop type difference coefficient is dynamically corrected to obtain a dynamic difference coefficient that updates in real time with the growth stage. ， is a dynamic difference coefficient, is an initial crop type difference coefficient, is a current growth stage temperature sensitivity threshold, is a crop full cycle average temperature sensitivity threshold; Centered on the blank pixel, a neighborhood range with a radius of R is defined, and the temperature data of all valid pixels within the neighborhood range are extracted; Introducing a spatial distance attenuation factor: ； wherein, is the spatial straight-line distance between the effective pixel and the blank pixel; The dynamic difference coefficient and the spatial distance attenuation factor are weighted and fused to obtain the comprehensive weight. The temperature value of each effective pixel is fitted with the corresponding comprehensive weight to calculate the initial predicted temperature of the blank pixel and fill it into the initial temperature matrix. With the three-dimensional coordinate system of the greenhouse X Axis, Y The axis is the matrix row and column dimension reference, and the pixel resolution of the matrix is unified. The initial temperature matrix after completion is dimensionally regularized to generate a global temperature matrix.

2. The greenhouse temperature control method according to claim 1, characterized in that, The steps of employing the RANSAC random sampling consensus algorithm, using sensor spatial coordinates and their corresponding temperature data as constraints, optimizing the transformation relationship between pixel coordinates and spatial coordinates, constructing a precise mapping model between sensor position and image pixels, and realizing the associative assignment of sensor temperature data to corresponding pixel regions include: Clean the sensor spatial coordinates and corresponding temperature data, and bind the sensor spatial coordinates with temperature values ​​to form constraint data in a preset format to establish a constraint dataset. Randomly select some samples from the constrained dataset, substitute them into the pixel coordinate and spatial coordinate transformation relationship to calculate the corresponding pixel coordinates, compare the calculated pixel coordinates with the actual pixel coordinates of the corresponding sensor area in the image, and obtain the projection error. Set an error threshold, classify samples with projection errors less than the threshold as inliers, and the rest as outliers. Iterate and repeat the process of extracting some samples, calculating errors, and filtering inliers. In each round, use the current set of inliers to reconstruct the pixel coordinates and spatial coordinates until the proportion of inliers is stable above the preset ratio. Output the final pixel coordinates and spatial coordinates transformation relationship. By combining the sensor coordinates with the final pixel coordinates and spatial coordinates transformation relationship, a bidirectional mapping channel is formed to construct a precise mapping model between the sensor position and image pixels. The bidirectional mapping channel includes a forward channel and a reverse channel. The forward channel outputs the corresponding image pixel coordinates by inputting the sensor's three-dimensional coordinates, while the reverse channel infers the corresponding three-dimensional spatial coordinates of the greenhouse by inputting the image pixel coordinates.

3. The greenhouse temperature control method according to claim 1, characterized in that, The steps for constructing a temperature influence factor model by combining external greenhouse environmental data include: Key factors in the greenhouse external environment data were screened and divided into dynamic factors and static factors. Dynamic factors include real-time wind speed, solar radiation intensity, and air humidity, while static factors include geographical location latitude and seasonal attributes. The mean value and mutation characteristics of dynamic factors are extracted over a unit time, and the static factors are converted into feature values ​​through standardization to construct an environmental feature set. Using environmental feature set as input and actual temperature change as output, a temperature influence factor model is obtained through regression algorithm training.

4. The greenhouse temperature control method according to claim 1, characterized in that, The step of regulating the temperature of each deviating area based on the spatial location, crop type, and temperature difference of each region until the temperature of the deviating area for greenhouse crop growth stabilizes within the corresponding suitable temperature threshold range includes: Based on the spatial location of each area in the greenhouse, the effective control area of ​​each temperature control device is marked to form a table of correspondence between devices and areas. At the same time, the crop types in each area are recorded, and the temperature sensitivity characteristics and suitable temperature requirements of different crops are queried and determined. Based on the magnitude of the temperature regulation difference, the deviation areas are divided into three categories: small difference, medium difference, and large difference. The regulation intensity is adjusted in combination with the crop type. Among them, the regulation is strengthened in the large difference area for warm-loving crops, and a mild regulation method is adopted in the small difference area for heat-resistant crops. Based on the equipment and area correspondence table, assign dedicated or nearby temperature control equipment to each off-center area until the temperature in the off-center area where greenhouse crops are growing stabilizes within the corresponding suitable temperature threshold range.

5. A greenhouse temperature control system, characterized in that, The system is used to implement the greenhouse temperature control method according to any one of claims 1 to 4, the system comprising: The acquisition module is used to acquire real-time temperature data of different crop growth areas collected by distributed temperature sensors inside the greenhouse, temperature field distribution images of the entire greenhouse collected by infrared thermal imaging devices, and environmental data outside the greenhouse collected by the environmental monitoring module. The fusion module is used to perform data fusion processing on real-time temperature data and temperature field distribution images. It establishes a global temperature matrix by mapping the pixel coordinates to the sensor position and constructs a temperature influence factor model by combining the external environmental data of the greenhouse. The calculation module is used to predict the temperature change trend of each area of ​​the greenhouse within a preset time period based on the temperature influence factor model, compare the current real-time temperature of each area with the corresponding suitable temperature threshold for the current crop, determine the temperature deviation area, and calculate the temperature control difference of each deviation area. The control module is used to regulate the temperature of each deviation area based on the spatial location, crop type, and temperature regulation difference of each area, until the temperature of the deviation area where the greenhouse crops are growing stabilizes within the corresponding suitable temperature threshold range.

6. A readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps of the method as described in any one of claims 1 to 4.

7. A computer device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the steps of the method as described in any one of claims 1 to 4.

Images