Multi-sensor fusion agricultural microclimate monitoring system, method and device

Abstract

The invention relates to an agricultural microclimate monitoring system, method and device based on multi-sensor fusion. The method comprises an agricultural microclimate prediction module used for obtaining basic climate forecast information and obtaining microclimate multi-mode prediction data; the crop growth period prediction module is used for acquiring crop growth state data and obtaining crop growth period prediction data; the operation instruction generation module is used for generating an agricultural operation control instruction; the microclimate data acquisition module is used for acquiring microclimate multi-modal sensing data; the operation growth period verification module is used for obtaining actual data of a crop growth period; and the agricultural microclimate correction module is used for generating agricultural microclimate monitoring result information. By adopting the system, multi-dimensional sensor data fusion can be realized, the prediction accuracy of the system is improved, and the adaptive capacity of the system is enhanced by constructing a closed-loop correction mechanism.

Description

Technical Field

[0001] The present invention belongs to the field of agricultural climate monitoring, and particularly relates to an agricultural microclimate monitoring system, method and device based on multi-sensor fusion. Background Art

[0002] With the improvement of the requirements for agricultural modernization and refined management, agricultural climate monitoring technology has gradually developed from manual observation to automated monitoring. Early technologies mainly relied on meteorological stations or sensors deployed at single points, and data collection was achieved through wired or simple wireless communication. Combined with basic statistical methods, the impact of climate parameters on crop growth was analyzed, mainly serving regional climate trend analysis.

[0003] In traditional agricultural climate monitoring and regulation, basic data is usually obtained by single-point fixed sensors, and data processing relies on manual summary and tabular analysis. Regulation strategies are mostly formulated based on historical average data, lacking the dynamic association between real-time climate data and crop growth status. Although it can meet the basic needs of agricultural production to a certain extent, there are significant deficiencies such as limited monitoring scope, serious data fragmentation, low efficiency of human-computer interaction, lagging regulation strategies, and low system integration.

[0004] In the prior art with the authorization announcement number of CN118758358B, a field microclimate automatic acquisition and reception system is provided. The system ensures the acquisition accuracy of meteorological data through automatic drainage of the rain sensor and multi-sensor integration; through the adaptive fixed structure of the support component, the stability of the system in a complex field environment is enhanced; the modular design supports flexible configuration and position adjustment of sensors, adapting to the monitoring needs of different crop growth stages, but there are obvious deficiencies in climate prediction, multi-sensor collaboration, and intelligent decision-making. Summary of the Invention

[0005] Based on this, in view of the above technical problems, it is necessary to provide an agricultural microclimate monitoring system, method and device based on multi-sensor fusion that can achieve multi-sensor dynamic fusion, intelligent decision-making for agricultural operations, and real-time closed-loop optimization.

[0006] In a first aspect, the present application provides an agricultural microclimate monitoring system based on multi-sensor fusion, including:

[0007] An agricultural microclimate prediction module, configured to obtain basic climate forecast information and input the basic climate forecast information into a microclimate sensing and prediction model to obtain microclimate multi-modal prediction data;

[0008] A crop growth period prediction module, configured to obtain crop growth status data and input the crop growth status data into a crop growth period prediction model to obtain crop growth period prediction data;

[0009] An operation instruction generation module, configured to generate an agricultural operation control instruction based on the prediction data of the crop generator and the multi-modal prediction data of the microclimate in combination with the preset agricultural operation rules;

[0010] A microclimate data acquisition module, configured to acquire multi-modal microclimate sensing data;

[0011] An operation growth period verification module, configured to input the crop growth state data into the crop growth period analysis model to obtain the actual data of the crop growth period;

[0012] An agricultural microclimate correction module, configured to perform error analysis on the multi-modal prediction data of the microclimate and the prediction data of the crop growth period according to the multi-modal microclimate sensing data and the actual data of the crop growth period, and generate agricultural microclimate monitoring result information, where the microclimate monitoring result information is used to correct the agricultural operation control instruction.

[0013] In one embodiment, the multi-modal microclimate sensing data includes natural multi-modal microclimate sensing data, and the natural multi-modal microclimate sensing data is used to characterize the multi-modal microclimate sensing data of the adjacent area outside the area affected by the artificial influence exerted based on the agricultural operation control instruction. The agricultural microclimate monitoring system with multi-sensor fusion further includes:

[0014] A prediction model optimization module, configured to feedback and optimize the microclimate sensing prediction model based on the natural multi-modal microclimate sensing data and the multi-modal prediction data of the microclimate.

[0015] In one embodiment, the prediction model optimization module includes:

[0016] A sensing weight analysis component, configured to generate the sensor weight of each sensor based on the accuracy information, aging information, and sampling frequency information of each sensor in the multi-modal sensor group for acquiring the natural multi-modal microclimate sensing data;

[0017] A weighted error analysis component, configured to generate the error data of each sensor based on the natural multi-modal microclimate sensing data and the corresponding multi-modal prediction data of the natural multi-modal microclimate sensing data, and generate the sensor weighted error data of each modality in combination with the sensor weight;

[0018] A weighted error determination component, configured to perform threshold determination on the sensor weighted error data of each modality based on the preset microclimate prediction error threshold of each modality, and generate the weighted error threshold determination result of each modality;

[0019] A prediction model update component, configured to update and optimize the microclimate sensing prediction model based on the natural multi-modal microclimate sensing data and the multi-modal prediction data of the modality corresponding to the weighted error threshold determination result if the weighted error threshold determination result exceeds the prediction error threshold;

[0020] Among them, the expression of the sensor weight is:

[0021]

[0022] In the formula, is the sensor weight of the i-th sensor in the j-th mode, is the sampling frequency weight coefficient of the j-th mode, is the sampling frequency of the i-th sensor in the j-th mode, is the average sampling frequency of the j-th mode, is the service life weight coefficient of the j-th mode, is the actual service life of the i-th sensor in the j-th mode, is the designed service life of the i-th sensor in the j-th mode, is the accuracy weight coefficient of the j-th mode, is the average accuracy of the j-th mode, is the accuracy of the i-th sensor in the j-th mode.

[0023] In one of the embodiments, the microclimate sensing prediction model is a neural network model constructed based on an improved residual convolutional network and an improved long short-term memory network;

[0024] The update and optimization loss function of the microclimate sensing prediction model includes an improved quantile weighted loss function. The update and optimization loss function is used to update and optimize the microclimate sensing prediction model based on the weighted error threshold determination result corresponding to the natural microclimate multimodal sensing data and microclimate multimodal prediction data of the mode. The expression of the improved quantile weighted loss function is:

[0025]

[0026] In the formula, L QW is the improved quantile weighted loss function, N is the number of samples, T is the time step, is the value of the quantile loss function at the quantile τ ν , y υ is the true value of the ν-th sample, is the prediction of the υ-th sample at the quantile τ ν .

[0027] In one of the embodiments, the agricultural microclimate correction module includes:

[0028] A regulation effect prediction component, configured to input an agricultural operation control instruction and microclimate multimodal prediction data into an agricultural operation control effect prediction model to generate agricultural operation control prediction result data;

[0029] A control error monitoring component for calculating the error between real-time multi-modal small climate sensing data and agricultural operation control prediction result data, and generating operation control error data;

[0030] A control error determination component for performing threshold determination on the operation control error data in combination with a preset operation control error threshold, and generating a control error alarm message if the operation control error data exceeds the operation control error threshold.

[0031] In one embodiment, the agricultural small climate correction module further includes:

[0032] A growth period error monitoring component for calculating the error between real-time actual data of the crop growth period and predicted data of the crop growth period, and generating crop growth period error data;

[0033] A growth period error determination component for performing threshold determination on the crop growth period error data in combination with a preset crop growth period time error threshold, and generating a model update instruction message for the crop growth period prediction model if the crop growth period error data exceeds the crop growth period time error threshold;

[0034] A growth period prediction model update component for updating the crop growth period prediction model based on the crop growth period error data.

[0035] In one embodiment, the input of the crop growth period prediction model further includes a historical multi-modal small climate sensing data set, and the historical multi-modal small climate sensing data set is constructed based on the multi-modal small climate sensing data.

[0036] In one embodiment, the multi-sensor fusion agricultural small climate monitoring system further includes:

[0037] A sensing data visualization module for mapping the multi-modal small climate sensing data, the multi-modal small climate prediction data, and the agricultural small climate monitoring result information to a farmland geographic information system to generate a multi-modal small climate sensing data display model;

[0038] A crop growth period visualization module for mapping the actual data of the crop growth period and the predicted data of the crop growth period to a farmland geographic information system to generate a crop growth period visualization display model.

[0039] In a second aspect, the present application further provides a multi-sensor fusion agricultural small climate monitoring method, including:

[0040] Obtaining basic climate forecast information and inputting the basic climate forecast information into a small climate sensing prediction model to obtain multi-modal small climate prediction data;

[0041] Obtain crop growth status data, and input the crop growth status data into a crop growth period prediction model to obtain crop growth period prediction data;

[0042] Based on the crop generator prediction data and the microclimate multimodal prediction data, combine the preset agricultural operation rules to generate an agricultural operation control instruction;

[0043] Obtain microclimate multimodal sensing data;

[0044] Input the crop growth status data into the crop growth period prediction model to obtain the actual crop growth period data;

[0045] Perform error analysis on the microclimate multimodal prediction data and the crop growth period prediction data according to the microclimate multimodal sensing data and the actual crop growth period data, and generate agricultural microclimate monitoring result information, which is used to correct the agricultural operation control instruction.

[0046] In a third aspect, the present application also provides a computer device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps of the method in the second aspect of the present application are implemented.

[0047] The above-mentioned agricultural microclimate monitoring system, method and device based on multi-sensor fusion integrate basic climate forecast information and crop growth status data through an agricultural microclimate prediction module and a crop growth period prediction module to form multimodal prediction data. Compared with a single sensor or an empirical model, this fusion mechanism can capture the dynamic relationship between climate factors and crop growth, significantly improving the accuracy of microclimate prediction and crop growth stage judgment; by constructing a "prediction - measurement - correction" closed loop through an operation growth period verification module and an agricultural microclimate correction module, it can automatically adapt to climate change and crop variety differences, avoiding the lag of traditional open-loop control, and making the regulation strategy more in line with real-time needs; through modular design, it can achieve compatibility expansion and flexible deployment, adapt to different farmland environments and crop types, and improve the versatility and flexibility of the system. Description of the Drawings

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following will briefly introduce the drawings required for use in the description of the embodiments or related technologies. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0049] Figure 1 It is a schematic structural diagram of an agricultural microclimate monitoring system based on multi-sensor fusion provided by an embodiment of the present application;

[0050] Figure 2 Schematic diagram of another multi-sensor fusion agricultural microclimate monitoring system provided by an embodiment of the present application;

[0051] Figure 3 Schematic diagram of a prediction model optimization module provided by an embodiment of the present application;

[0052] Figure 4 Schematic diagram of an agricultural microclimate correction module provided by an embodiment of the present application;

[0053] Figure 5 Schematic diagram of another agricultural microclimate correction module provided by an embodiment of the present application;

[0054] Figure 6 Schematic diagram of a multi-sensor fusion agricultural microclimate monitoring method provided by an embodiment of the present application. Detailed implementation manners

[0055] In order to make the objectives, technical solutions and advantages of the present application clearer and more understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0056] In one embodiment, as Figure 1 shown, a multi-sensor fusion agricultural microclimate monitoring system 100 is provided. In this embodiment, the system includes:

[0057] An agricultural microclimate prediction module 101, which can be used to obtain basic climate forecast information and input the basic climate forecast information into a microclimate sensing prediction model to obtain microclimate multi-modal prediction data.

[0058] Illustratively, since climate prediction has high requirements for both algorithm design and computing power, public existing basic climate forecast information can be used for migration analysis, extracting the correlation features between basic climate forecast data and agricultural microclimate data, and predicting future agricultural microclimate data based on historical agricultural microclimate data and basic climate forecast information.

[0059] Optionally, the basic climate forecast information may include, but is not limited to, weather forecast information obtained from a public platform.

[0060] Optionally, the microclimate multi-modal data may include, but is not limited to, temperature mode, humidity mode, light mode, wind force mode and precipitation mode.

[0061] The crop growth period prediction module 102 can be used to obtain crop growth status data and input the crop growth status data into a crop growth period prediction model to obtain crop growth period prediction data.

[0062] Optionally, the crop growth status data may include, but is not limited to, drone image data and fixed-position image data.

[0063] Optionally, the crop growth period may include, but is not limited to, the tillering stage, jointing stage, heading stage, flowering and pollination stage, filling stage, maturity stage, tasseling stage, small bellmouth stage, large bellmouth stage, flowering and fruit setting stage, and fruit development stage.

[0064] The operation instruction generation module 103 can be used to generate an agricultural operation control instruction based on the crop generator prediction data and the microclimate multimodal prediction data in combination with preset agricultural operation rules.

[0065] Illustratively, the agricultural operation control instruction may include, but is not limited to, drainage, irrigation, fertilization, supplementary lighting, film mulching, shading, soil improvement, ventilation, intercropping and relay cropping, timely harvesting, and weed removal.

[0066] The microclimate data acquisition module 104 can be used to obtain microclimate multimodal sensing data.

[0067] The operation growth period verification module 105 can be used to input the crop growth status data into a crop growth period analysis model to obtain the actual crop growth period data.

[0068] Illustratively, compared with climate prediction, the computing power required for crop stage prediction is less, and the crop stage can be predicted based on the existing crop stage status data.

[0069] The agricultural microclimate correction module 106 can be used to perform error analysis on the microclimate multimodal prediction data and the crop growth period prediction data according to the microclimate multimodal sensing data and the actual crop growth period data, generate agricultural microclimate monitoring result information, and the microclimate monitoring result information is used to correct the agricultural operation control instruction.

[0070] In the above-mentioned agricultural microclimate monitoring system with multi-sensor fusion, by fusing climate sensors and crop sensors collected by multiple sensors, it can provide comprehensive and accurate microclimate monitoring. It can not only cover environmental factors but also involve the real-time growth status of crops, making the monitoring data more representative. By constructing a microclimate sensing prediction model and a crop growth period prediction model, it can analyze historical data and real-time data to predict future microclimate conditions and crop growth stages, providing forward-looking guidance for agricultural production. Through the operation instruction generation module based on the prediction data and preset agricultural operation rules, it can automatically generate accurate agricultural operation control instructions, enabling operations such as irrigation, fertilization, light supplementation, film covering, and sunshading at the appropriate time, thereby improving the efficiency and quality of agricultural production. By comparing the microclimate multi-modal sensing data with the actual data of the crop growth period, conducting error analysis on the prediction data, and generating agricultural microclimate monitoring result information, it can correct the agricultural operation control instructions, thereby ensuring the reliability of prediction and decision-making.

[0071] In an alternative embodiment of the present application, the microclimate multi-modal sensing data includes natural microclimate multi-modal sensing data, which is used to represent the microclimate multi-modal sensing data in the adjacent area outside the area affected by human influence based on the agricultural operation control instructions. Please refer to Figure 2 , the agricultural microclimate monitoring system 100 with multi-sensor fusion may further include:

[0072] A prediction model optimization module 207, which can be used to optimize the microclimate sensing prediction model based on the natural microclimate multi-modal sensing data and the microclimate multi-modal prediction data feedback.

[0073] In the above-mentioned agricultural microclimate monitoring system with multi-sensor fusion, by optimizing the microclimate sensing prediction model based on the natural microclimate multi-modal sensing data and the microclimate multi-modal prediction data feedback, it can improve the monitoring accuracy, data prediction ability, decision-making effectiveness and scientificity, and the reliability of the system of the agricultural microclimate monitoring system with multi-sensor fusion, thereby promoting the sustainable development of agriculture.

[0074] In an alternative embodiment of the present application, as Figure 3 shown, the prediction model optimization module 207 may include:

[0075] A sensing weight analysis component 301, which can be used to generate the sensor weights of each sensor based on the accuracy information, aging information, and sampling frequency information of each sensor in the multi-modal sensor group for obtaining the natural microclimate multi-modal sensing data.

[0076] The weighted error analysis component 302 can be used to generate error data for each sensor based on the natural microclimate multimodal sensing data and the microclimate multimodal prediction data corresponding to the natural microclimate multimodal sensing data, and generate the sensor weighted error data for each modality by combining the sensor weights.

[0077] The weighted error determination component 303 can be used to perform threshold determination on the sensor weighted error data for each modality based on the preset microclimate prediction error thresholds for each modality, and generate the weighted error threshold determination results for each modality.

[0078] The prediction model update component 304 can be used to update and optimize the microclimate sensing prediction model based on the natural microclimate multimodal sensing data and the microclimate multimodal prediction data of the modality corresponding to the weighted error threshold determination result if the weighted error threshold determination result exceeds the prediction error threshold.

[0079] Among them, the expression of the sensor weight can be:

[0080]

[0081] In the formula, is the sensor weight of the i-th sensor in the j-th modality, is the sampling frequency weight coefficient of the j-th modality, is the sampling frequency of the i-th sensor in the j-th modality, is the average sampling frequency of the j-th modality, is the service life weight coefficient of the j-th modality, is the actual service life of the i-th sensor in the j-th modality, is the designed service life of the i-th sensor in the j-th modality, is the accuracy weight coefficient of the j-th modality, is the average accuracy of the j-th modality, is the accuracy of the i-th sensor in the j-th modality.

[0082] In the above agricultural microclimate monitoring system with multi-sensor fusion, by generating sensor weights according to the accuracy information, aging information, and sampling frequency information of each sensor in the multimodal sensor group, it can ensure that sensors with high accuracy, high sampling frequency, and high reliability have a greater weight in data fusion, thereby improving the effectiveness and scientific nature of data fusion; through the prediction model update component, the system can continuously optimize and improve the prediction model, continuously enhance the prediction accuracy and reliability, and further ensure the adaptability of the system to the complex and changeable farmland environment, and reduce the cost of system maintenance and update.

[0083] In an alternative embodiment of the present application, the microclimate sensing prediction model is a neural network model constructed based on an improved residual convolutional network and an improved long short-term memory network.

[0084] The update and optimization loss function of the microclimate sensing prediction model includes an improved quantile weighted loss function. The update and optimization loss function is used to update and optimize the microclimate sensing prediction model based on the natural microclimate multi-modal sensing data and microclimate multi-modal prediction data corresponding to the modal determined by the weighted error threshold. The expression of the improved quantile weighted loss function is:

[0085]

[0086] In the formula, L QW is the improved quantile weighted loss function, N is the number of samples, T is the time step, is the value of the quantile loss function at the quantile τ ν , y υ is the true value of the υ-th sample, is the prediction of the υ-th sample at the quantile τ v .

[0087] In the above multi-sensor fusion agricultural microclimate monitoring system, by combining the quantile loss function and the sensor weights to construct an improved quantile weighted loss function, the prediction error can be quantified more accurately. It can not only consider the deviation between the predicted value and the true value, but also combine the reliability of the sensors, so that the error evaluation can be more accurate, and the pertinence and effectiveness of model optimization can be improved; the microclimate sensing prediction model constructed based on the improved residual convolutional network and the improved long short-term memory network can better adapt to the complex and changeable farmland environment. Among them, the residual convolutional network helps to capture spatial features, while the long short-term memory network can effectively process time series data. The combination of the two can improve the ability of the model to process multi-sensor data.

[0088] In an alternative embodiment of the present application, as Figure 4 shown, the agricultural microclimate correction module 106 may include:

[0089] The regulation effect prediction component 401 can be used to input the agricultural operation control instruction and the microclimate multi-modal prediction data into the agricultural operation control effect prediction model to generate the agricultural operation control prediction result data.

[0090] Optionally, the agricultural operation control effect prediction model can predict the control effect of the agricultural operation control instruction under the microclimate conditions corresponding to the microclimate multi-modal prediction data.

[0091] The control error monitoring component 402 can be used to calculate the error between the real-time microclimate multi-modal sensing data and the agricultural operation control prediction result data, and generate operation control error data.

[0092] The control error determination component 403 can be used to perform threshold determination on the operation control error data in combination with a preset operation control error threshold. If the operation control error data exceeds the operation control error threshold, it generates a control error alarm message.

[0093] In the above multi-sensor fusion agricultural microclimate monitoring system, by predicting the implementation effects of different operation measures, it is possible to achieve reasonable resource allocation, improve the agricultural resource utilization efficiency, and reduce the agricultural production cost; through the self-correction and early warning mechanisms, it can ensure the stability and reliability of the system during long-term use, and reduce the costs of system maintenance and update.

[0094] In an alternative embodiment of the present application, please refer to Figure 5 , the agricultural microclimate correction module 106 may further include:

[0095] The growth period error monitoring component 504 can be used to calculate the error between the real-time actual data of the crop growth period and the predicted data of the crop growth period, and generate crop growth period error data.

[0096] The growth period error determination component 505 can be used to perform threshold determination on the crop growth period error data in combination with a preset crop growth period time error threshold. If the crop growth period error data exceeds the crop growth period time error threshold, it generates a model update instruction message for the crop growth period prediction model.

[0097] The growth period prediction model update component 506 can be used to update the crop growth period prediction model based on the crop growth period error data.

[0098] In the above multi-sensor fusion agricultural microclimate monitoring system, by generating crop growth period error data to construct a real-time monitoring mechanism, it can timely detect the deviation between crop growth and prediction, and thus can realize the dynamic adjustment of the crop growth period prediction model.

[0099] In an alternative embodiment of the present application, the input of the crop growth period prediction model may further include a historical microclimate multi-modal sensing data set, and the historical microclimate multi-modal sensing data set may be a data set constructed based on the microclimate multi-modal sensing data.

[0100] In an alternative embodiment of the present application, please refer to Figure 2 , the multi-sensor fusion agricultural microclimate monitoring system 100 may further include:

[0101] The sensing data visualization module 208 can be used to map the microclimate multimodal sensing data, microclimate multimodal prediction data, and agricultural microclimate monitoring result information to the farmland geographic information system to generate a microclimate multimodal sensing data display model.

[0102] The crop growth period visualization module 209 can be used to map the actual data of the crop growth period and the predicted data of the crop growth period to the farmland geographic information system to generate a crop growth period visualization display model.

[0103] In the above agricultural microclimate monitoring system with multi-sensor fusion, by generating an intuitive microclimate multimodal sensing data display model, it can efficiently and clearly display the microclimate conditions and change trends in different areas of the farmland. Furthermore, it can quickly identify the areas with abnormal microclimate conditions and take targeted agricultural operation measures in a timely manner; by generating a crop growth period visualization display model, it can intuitively display the growth stages of crops in different areas, which helps to understand the uniformity and consistency of crop growth.

[0104] Based on the same inventive concept, the embodiment of the present application also provides a multi-sensor fusion agricultural microclimate monitoring method for implementing the above-mentioned agricultural microclimate monitoring system with multi-sensor fusion. The solution provided by this system to solve problems is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the multi-sensor fusion agricultural microclimate monitoring method provided below can refer to the limitations on the multi-sensor fusion agricultural microclimate monitoring method in the above text, and will not be repeated here.

[0105] In an exemplary embodiment of the present application, as Figure 6 shown, a multi-sensor fusion agricultural microclimate monitoring method is provided, including:

[0106] Step S601, obtain the basic climate forecast information, and input the basic climate forecast information into the microclimate sensing and prediction model to obtain microclimate multimodal prediction data.

[0107] Step S602, obtain the crop growth status data, and input the crop growth status data into the crop growth period prediction model to obtain the crop growth period prediction data.

[0108] Step S603, based on the crop generator prediction data and the microclimate multimodal prediction data, combine the preset agricultural operation rules to generate an agricultural operation control instruction.

[0109] Step S604, obtain the microclimate multimodal sensing data.

[0110] Step S605, input the crop growth status data into the crop growth period prediction model to obtain the actual data of the crop growth period.

[0111] Step S606: Perform error analysis on the small climate multi-modal prediction data and the crop growth period prediction data according to the small climate multi-modal sensing data and the actual data of the crop growth period, and generate agricultural small climate monitoring result information.

[0112] Optionally, the small climate monitoring result information can be used to correct the agricultural operation control instruction.

[0113] In an optional embodiment of the present application, the small climate multi-modal sensing data includes natural small climate multi-modal sensing data, and the natural small climate multi-modal sensing data is used to characterize the small climate multi-modal sensing data in the adjacent area outside the area affected by the artificial influence exerted based on the agricultural operation control instruction. The multi-sensor fusion agricultural small climate monitoring method further includes:

[0114] Feedback and optimize the small climate sensing prediction model based on the natural small climate multi-modal sensing data and the small climate multi-modal prediction data.

[0115] In an optional embodiment of the present application, feedback and optimizing the small climate sensing prediction model based on the natural small climate multi-modal sensing data and the small climate multi-modal prediction data may include:

[0116] Generate the sensor weight of each sensor based on the accuracy information, aging information, and sampling frequency information of each sensor in the multi-modal sensor group for obtaining the natural small climate multi-modal sensing data.

[0117] Generate the error data of each sensor based on the natural small climate multi-modal sensing data and the small climate multi-modal prediction data corresponding to the natural small climate multi-modal sensing data, and generate the sensor weighted error data of each modality in combination with the sensor weight.

[0118] Perform threshold determination on the sensor weighted error data of each modality based on the preset small climate prediction error threshold of each modality, and generate the weighted error threshold determination result of each modality.

[0119] If the weighted error threshold determination result is that it exceeds the prediction error threshold, update and optimize the small climate sensing prediction model based on the natural small climate multi-modal sensing data and the small climate multi-modal prediction data of the modality corresponding to the weighted error threshold determination result.

[0120] In an optional embodiment of the present application, performing error analysis on the small climate multi-modal prediction data and the crop growth period prediction data according to the small climate multi-modal sensing data and the actual data of the crop growth period, and generating agricultural small climate monitoring result information includes:

[0121] Input the agricultural operation control instruction and the small climate multi-modal prediction data into the agricultural operation control effect prediction model to generate agricultural operation control prediction result data.

[0122] Calculate the error between the real-time microclimate multimodal sensing data and the agricultural operation control prediction result data, and generate operation control error data.

[0123] Combine the preset operation control error threshold to perform threshold determination on the operation control error data. If the operation control error data exceeds the operation control error threshold, generate control error alarm information.

[0124] In an alternative embodiment of the present application, when performing error analysis on the microclimate multimodal prediction data and the crop growth period prediction data according to the microclimate multimodal sensing data and the actual data of the crop growth period to generate agricultural microclimate monitoring result information, it further includes:

[0125] Calculate the error between the real-time actual data of the crop growth period and the predicted data of the crop growth period, and generate crop growth period error data.

[0126] Combine the preset crop growth period time error threshold to perform threshold determination on the crop growth period error data. If the crop growth period error data exceeds the crop growth period time error threshold, generate model update instruction information for the crop growth period prediction model.

[0127] Update the crop growth period prediction model based on the crop growth period error data.

[0128] In an alternative embodiment of the present application, the agricultural microclimate monitoring method of multi-sensor fusion further includes:

[0129] Map the microclimate multimodal sensing data, the microclimate multimodal prediction data, and the agricultural microclimate monitoring result information to the farmland geographic information system to generate a microclimate multimodal sensing data display model.

[0130] Map the actual data of the crop growth period and the predicted data of the crop growth period to the farmland geographic information system to generate a crop growth period visualization display model.

[0131] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are displayed in sequence according to the arrows, these steps do not necessarily need to be executed in the order indicated by the arrows. Unless there is a clear description in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages do not necessarily need to be executed at the same time, but can be executed at different times. The execution order of these steps or stages does not necessarily need to be sequential, but can be executed alternately or alternately with at least a part of other steps or steps in other steps.

[0132] In one embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps of a multi-sensor fusion-based agricultural microclimate monitoring method as described above are implemented.

[0133] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0134] For the device embodiment, since it basically corresponds to the method embodiment, the relevant parts can be referred to the partial description of the method embodiment. The device embodiments described above are merely illustrative. The components described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the present disclosure solution. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0135] The above embodiments only represent several implementation manners of the embodiments of the present application. The description is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of the embodiments of the application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the embodiments of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the embodiments of the present application.

Claims

1. An agricultural microclimate monitoring system with multi-sensor fusion, characterized in that, The system includes: An agricultural microclimate prediction module, configured to obtain basic climate forecast information and input the basic climate forecast information into a microclimate sensing prediction model to obtain microclimate multimodal prediction data; A crop growth period prediction module, configured to obtain crop growth status data and input the crop growth status data into a crop growth period prediction model to obtain crop growth period prediction data; An operation instruction generation module, configured to generate an agricultural operation control instruction based on the crop generator prediction data and the microclimate multimodal prediction data in combination with a preset agricultural operation rule; A microclimate data acquisition module, configured to obtain microclimate multimodal sensing data; An operation growth period verification module, configured to input the crop growth status data into a crop growth period analysis model to obtain actual crop growth period data; An agricultural microclimate correction module, configured to perform error analysis on the microclimate multimodal prediction data and the crop growth period prediction data according to the microclimate multimodal sensing data and the actual crop growth period data, and generate agricultural microclimate monitoring result information, where the microclimate monitoring result information is used to correct the agricultural operation control instruction.

2. The system according to claim 1, wherein The microclimate multimodal sensing data includes natural microclimate multimodal sensing data, and the natural microclimate multimodal sensing data is used to represent the microclimate multimodal sensing data in an adjacent area outside the area affected by human influence based on the agricultural operation control instruction. The system further includes: A prediction model optimization module, configured to feedback and optimize the microclimate sensing prediction model based on the natural microclimate multimodal sensing data and the microclimate multimodal prediction data.

3. The system according to claim 2, wherein The prediction model optimization module includes: A sensing weight analysis component, configured to generate a sensor weight for each of the sensors in a multimodal sensor group for obtaining the natural microclimate multimodal sensing data based on the accuracy information, aging information, and sampling frequency information of each sensor; A weighted error analysis component, configured to generate error data for each of the sensors based on the natural microclimate multimodal sensing data and the corresponding microclimate multimodal prediction data of the natural microclimate multimodal sensing data, and generate sensor weighted error data for each modality in combination with the sensor weight; A weighted error determination component, configured to perform threshold determination on the sensor weighted error data for each of the modalities based on a preset microclimate prediction error threshold for each of the modalities to generate a weighted error threshold determination result for each of the modalities; A prediction model update component, configured to update and optimize the microclimate sensing prediction model based on the natural microclimate multimodal sensing data and the microclimate multimodal prediction data for the modality corresponding to the weighted error threshold determination result if the weighted error threshold determination result exceeds the prediction error threshold; Wherein, the expression of the sensor weight is: Wherein, is the sensor weight of the i-th sensor in the j-th mode, is the sampling frequency weight coefficient of the j-th mode, is the sampling frequency of the i-th sensor in the j-th mode, is the average sampling frequency of the j-th mode, is the service life weight coefficient of the j-th mode, is the actual service life of the i-th sensor in the j-th mode, is the designed service life of the i-th sensor in the j-th mode, is the accuracy weight coefficient of the j-th mode, is the average accuracy of the j-th mode, is the accuracy of the i-th sensor in the j-th mode.

4. The system according to claim 3, wherein: The microclimate sensing prediction model is a neural network model constructed based on an improved residual convolutional network and an improved long short-term memory network; The update and optimization loss function of the microclimate sensing prediction model includes an improved quantile weighted loss function. The update and optimization loss function is used to update and optimize the microclimate sensing prediction model based on the natural microclimate multimodal sensing data and the microclimate multimodal prediction data corresponding to the determination result of the weighted error threshold. The expression of the improved quantile weighted loss function is: Where L QW is the improved quantile weighted loss function, N is the number of samples, T is the time step, is the value of the quantile loss function at the quantile τ ν , y υ is the true value of the υ-th sample, is the prediction of the υ-th sample at the quantile τ ν .

5. The system according to claim 1, wherein The agricultural microclimate correction module includes: A regulation effect prediction component, which is used to input the agricultural operation control instruction and the microclimate multimodal prediction data into the agricultural operation control effect prediction model to generate agricultural operation control prediction result data; A control error monitoring component, which is used to calculate the error between the real-time microclimate multimodal sensing data and the agricultural operation control prediction result data to generate operation control error data; A control error determination component, which is used to perform threshold determination on the operation control error data in combination with a preset operation control error threshold. If the operation control error data exceeds the operation control error threshold, it generates a control error alarm message.

6. The system according to claim 5, wherein The agricultural microclimate correction module further includes: A growth period error monitoring component, which is used to calculate the error between the real-time actual data of the crop growth period and the predicted data of the crop growth period to generate crop growth period error data; A growth period error determination component, which is used to perform threshold determination on the crop growth period error data in combination with a preset crop growth period time error threshold. If the crop growth period error data exceeds the crop growth period time error threshold, it generates a model update instruction message for the crop growth period prediction model; A growth period prediction model update component, which is used to update the crop growth period prediction model based on the crop growth period error data.

7. The system according to claim 6, wherein: The input of the crop growth period prediction model further includes a historical microclimate multimodal sensing data set, and the historical microclimate multimodal sensing data set is constructed based on the microclimate multimodal sensing data.

8. The system according to any one of claims 1 to 7, characterized in that, The system further includes: A sensing data visualization module, which is used to map the microclimate multimodal sensing data, the microclimate multimodal prediction data, and the agricultural microclimate monitoring result information to the farmland geographic information system to generate a microclimate multimodal sensing data display model; A crop growth period visualization module, which is used to map the actual data of the crop growth period and the predicted data of the crop growth period to the farmland geographic information system to generate a crop growth period visualization display model.

9. A multi-sensor fusion-based agricultural microclimate monitoring method, characterized in that, The method includes: Obtaining basic climate forecast information and inputting the basic climate forecast information into the microclimate sensing prediction model to obtain microclimate multimodal prediction data; Obtaining crop growth status data and inputting the crop growth status data into the crop growth period prediction model to obtain crop growth period prediction data; Generating an agricultural operation control instruction based on the crop generator prediction data and the microclimate multimodal prediction data in combination with a preset agricultural operation rule; Obtaining microclimate multimodal sensing data; Input the crop growth status data into the crop growth period prediction model to obtain the actual crop growth period data; Perform error analysis on the small climate multimodal prediction data and the crop growth period prediction data according to the small climate multimodal sensing data and the actual crop growth period data, and generate agricultural small climate monitoring result information, which is used to correct the agricultural operation control instruction.

10. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, the method described in claim 9 is implemented.

Images