Crop growth prediction system and method based on digital twinning

Abstract

The application discloses a crop growth prediction system and method based on digital twinning, and belongs to the technical field of agricultural planting and plant factory. The system collects space-air-ground multi-source data of crops through a data acquisition layer; performs feature extraction processing on the space-air-ground multi-source data through a data processing layer, and constructs a multi-dimensional growth parameter model based on crop characteristic data; constructs a crop growth digital twinning model based on the multi-dimensional growth parameter model, crop historical data and regional climate characteristic information through a digital twinning engine layer; simulates the growth process and response of crops under different planting strategies through an intelligent decision layer, and generates planting decision suggestions according to the simulation results; and visually displays the virtual growth state of crops based on the simulation results through a visual display layer. The system can effectively improve the efficiency and accuracy of crop growth management.

Description

Technical Field

[0001] This application belongs to the field of agricultural planting and plant factory technology, specifically relating to a crop growth prediction system and method based on digital twins. Background Technology

[0002] With the rapid development of science and technology, agricultural production methods are shifting from traditional experience-based management to intelligent management. How to utilize intelligent technology to achieve crop growth management and prediction is an important research direction.

[0003] Traditional crop management methods rely primarily on manual inspections, experience-based judgments, and handwritten records, making it difficult to achieve precise control over the crop growth environment and optimal resource allocation. With the development of IoT and sensor technologies, remote sensing and IoT data collection are gradually being incorporated into crop production management. However, current methods still rely on single-source data collection, which cannot comprehensively reflect the actual growth status of crops in complex environments, resulting in insufficient and inaccurate monitoring information. Existing plant factory technologies typically use fixed or empirical parameters, leading to limited perception of growth status and an inability to provide targeted planting decisions based on the differentiated characteristics of crops. This results in low efficiency and poor accuracy in crop growth management.

[0004] Therefore, there is a need for a system and method for more efficient and accurate prediction and management of crop growth. Summary of the Invention

[0005] The purpose of this application is to provide a crop growth prediction system and method based on digital twins, which can solve the problems of low efficiency and poor accuracy in crop growth management in related technologies.

[0006] To solve the above-mentioned technical problems, this application is implemented as follows: In a first aspect, this application provides a crop growth prediction system based on digital twins, the system comprising: The data acquisition layer is used to collect multi-source data on crops from air, space, and ground. The data processing layer is used to perform feature extraction processing based on the multi-source data from air, space, and ground to obtain crop feature data, and to construct a multi-dimensional growth parameter model representing the growth status of crops based on the crop feature data. The digital twin engine layer is used to construct a digital twin model of crop growth based on the multidimensional growth parameter model, historical air-space-ground multi-source data of crops, and regional climate characteristic information, and to predict the growth status of crops based on the digital twin model of crop growth. The intelligent decision-making layer is used to simulate the growth process and response of crops under different planting strategies based on the crop growth digital twin model, obtain simulation results, and generate planting decision suggestions based on the simulation results. The planting strategies include irrigation strategies, fertilization schemes, and sowing schemes under different weather conditions. The visualization layer is used to visualize the virtual growth status of crops based on the simulation results.

[0007] Secondly, this application provides a crop growth prediction method based on digital twins, the method comprising: Acquire multi-source data on crops from air, space, and ground; Based on the multi-source data from air, space, and ground, feature extraction processing is performed to obtain crop feature data, and a multi-dimensional growth parameter model characterizing the growth status of crops is constructed based on the crop feature data. Based on the multidimensional growth parameter model, historical air-space-ground multi-source data of crops, and regional climate characteristic information, a digital twin model of crop growth is constructed, and the growth status of crops is predicted based on the digital twin model of crop growth. Based on the aforementioned digital twin model of crop growth, the growth process and response of crops under different planting strategies are simulated to obtain simulation results. Based on the simulation results, planting decision suggestions are generated. The planting strategies include irrigation strategies, fertilization schemes, and sowing schemes under different weather conditions. The virtual growth status of crops is visualized based on the simulation results.

[0008] Thirdly, this application provides a sugar beet growth prediction system based on digital twins, the system comprising: The data acquisition layer is used to collect multi-source data on sugar beets from air, space, and ground. The data processing layer is used to perform feature extraction processing based on the multi-source data from air, space, and ground to obtain beet feature data, and to construct a multi-dimensional growth parameter model representing the growth state of beet based on the beet feature data. The multi-dimensional growth parameter model includes: morphological development model, physiological and biochemical model and yield formation model. The digital twin engine layer is used to construct a digital twin model of sugar beet growth based on the multidimensional growth parameter model, historical air-space-ground multi-source data of sugar beets, and regional climate characteristic information, and to predict the growth status of sugar beets based on the digital twin model of sugar beet growth. The intelligent decision-making layer is used to simulate the growth process and response of sugar beets under different planting strategies based on the digital twin model of sugar beet growth, obtain simulation results, and generate planting decision suggestions based on the simulation results. The planting strategies include irrigation strategies, fertilization schemes and sowing schemes under different weather conditions. The visualization layer is used to visualize the virtual growth state of sugar beets based on the simulation results.

[0009] Fourthly, this application provides an electronic device, including a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, performs the following steps: Acquire multi-source data on crops from air, space, and ground; Based on the multi-source data from air, space, and ground, feature extraction processing is performed to obtain crop feature data, and a multi-dimensional growth parameter model characterizing the growth status of crops is constructed based on the crop feature data. Based on the multidimensional growth parameter model, historical air-space-ground multi-source data of crops, and regional climate characteristic information, a digital twin model of crop growth is constructed, and the growth status of crops is predicted based on the digital twin model of crop growth. Based on the aforementioned digital twin model of crop growth, the growth process and response of crops under different planting strategies are simulated to obtain simulation results. Based on the simulation results, planting decision suggestions are generated. The planting strategies include irrigation strategies, fertilization schemes, and sowing schemes under different weather conditions. The virtual growth status of crops is visualized based on the simulation results.

[0010] The technical solution provided in this application may include the following beneficial effects: This application provides a crop growth prediction system based on digital twins. The system mainly includes: a data acquisition layer, a data processing layer, a digital twin engine layer, an intelligent decision-making layer, and a visualization layer. The data acquisition layer collects multi-source, heterogeneous data from space, air, and ground, including multispectral images of crops, microenvironmental parameters, and biochemical indicators. This facilitates comprehensive monitoring and dynamic analysis of the crop growth environment and provides a basis for establishing multi-dimensional growth parameter models. The data processing layer extracts features from the space, air, and ground data in both temporal and spatial dimensions and constructs multi-dimensional growth parameter models, enabling a comprehensive and accurate reflection of crop growth under complex environments. The digital twin engine layer, with its constructed digital twin model of crop growth, and the intelligent decision-making layer's "hypothesis-verification" experimental function, simulates the growth state of crops under different planting strategies. This allows for continuous adjustment and optimization, achieving accurate simulation, prediction, and optimization of the actual planting process. This results in precision planting and yield optimization, significantly reducing decision-making risks, improving the efficiency and rationality of planting decisions, and ultimately enhancing the efficiency and accuracy of crop production management.

[0011] Furthermore, the data processing layer in this application extracts features based on temporal and spatial characteristics, and the crop digital twin model is constructed based on historical air-space-ground multi-source data and regional climate characteristic information. Through time series analysis and deep learning methods, it can predict the future growth trend and potential problems of crops, and also provide important early warnings for scientifically planning planting schemes and scientifically responding to emergencies in the crop growth process, thereby further improving the efficiency of agricultural production management. Attached Figure Description

[0012] Figure 1 This is a schematic diagram of the structure of a crop growth prediction system based on digital twins disclosed in an embodiment of this application; Figure 2 This is a flowchart illustrating a crop growth prediction method based on digital twins disclosed in an embodiment of this application; Figure 3 This is a schematic diagram of the structure of an electronic device disclosed in an embodiment of this application.

[0013] Figure labeling: 110: Data acquisition layer, 120: Data processing layer, 130: Digital twin engine layer, 140: Intelligent decision-making layer, 150: Visualization layer, 301: Processor, 302: Memory, 303: Power supply, 304: Wired or wireless network interface, 305: Input / output interface, 306: Keyboard. Detailed Implementation

[0014] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0015] The technical solutions disclosed in the various embodiments of this application are described in detail below with reference to the accompanying drawings.

[0016] Example 1 like Figure 1 As shown in the embodiments of this specification, a crop growth prediction system based on digital twins is provided. This crop growth prediction system includes: Data acquisition layer 110 is used to collect multi-source data on crops from air, space, and ground.

[0017] The multi-source data of crops includes: satellite remote sensing data, multispectral image data collected by drones equipped with multispectral cameras, temperature data collected by drones equipped with thermal infrared cameras, ground environmental parameters (such as soil moisture, water content, light intensity, etc.), and biochemical index data collected manually (such as chlorophyll content, nitrogen, phosphorus and potassium content, protein content, soluble sugar or starch content, etc.).

[0018] In this embodiment, the data acquisition layer can be implemented in various ways. The following is one implementation method, which includes: a multispectral image acquisition module, a microenvironment parameter acquisition module, a biochemical index acquisition module, and an integrated air-space-ground sensor network.

[0019] The system comprises several modules: a multispectral image acquisition module for acquiring remote multispectral image data of crops using satellite and drone-mounted multispectral cameras; a microenvironment parameter acquisition module for collecting environmental parameters such as soil moisture, temperature, light intensity, and CO2 concentration; a soil moisture sensor for soil moisture, a temperature sensor for ambient temperature, a light intensity sensor for light intensity, and a CO2 concentration sensor for CO2 concentration; and a biochemical indicator acquisition module for collecting biochemical indicators of crops, including chlorophyll content and protein content. These indicators can be obtained through manual sampling, periodically by using a chlorophyll meter to collect chlorophyll content, and measuring protein content using a protein content testing reagent. A space-air-ground integrated sensor network receives multi-source data from the multispectral image acquisition module, the microenvironment parameter acquisition module, and the biochemical indicator acquisition module via aggregation nodes, smart gateways, and cloud servers, and transmits this multi-source data to the data processing layer.

[0020] By setting up the above modules, the data acquisition layer can achieve comprehensive monitoring of the crop growth environment, avoiding the problem of inaccurate growth prediction caused by single information collection.

[0021] See also Figure 1 It is known that the system includes a data processing layer 120, which is used to perform feature extraction processing based on multi-source data from air, space, and ground to obtain crop feature data, and to construct a multi-dimensional growth parameter model representing the growth status of crops based on the crop feature data.

[0022] In implementation, the data processing layer can include a data preprocessing module and a data analysis module. The data preprocessing module is used to preprocess multi-source data (space, air, and ground) to obtain preprocessed data. Preprocessing includes standardization, normalization, missing value handling, data cleaning, data correction, and data transformation. The data analysis module is used to perform feature extraction on the preprocessed data to obtain crop feature data, and then performs pattern recognition processing on the crop feature data to obtain a multi-dimensional growth parameter model representing the crop's growth status. Feature extraction includes temporal feature extraction and spatial feature extraction.

[0023] In this embodiment, sugar beets are used as an example of crops. Spatial features include leaf texture features, leaf shape, leaf size, leaf length, leaf width, etc. Temporal features include growth rate, growth cycle, etc. For example, when there are 2 leaves, it represents which growth cycle the crop is in, and when there are 4 leaves, it represents which growth cycle the crop is in, etc.

[0024] In implementation, the data analysis module can use deep learning algorithms to extract temporal and spatial features from the preprocessed data to obtain crop feature data in the form of multidimensional feature vectors. Then, pattern recognition processing is performed on the crop feature data and its corresponding annotation information to obtain a multidimensional growth parameter model.

[0025] Deep learning algorithms for feature extraction can be: using recurrent neural networks or Transformer / self-attention mechanisms to extract temporal features, and using convolutional neural networks, graph neural networks or ViT visual transformers to extract spatial features.

[0026] The process of obtaining a multidimensional growth parameter model by performing pattern recognition processing on crop feature data can be carried out using a supervised learning algorithm. That is, using crop feature data and corresponding annotation information as training samples, performing pattern recognition processing on the crop feature data and corresponding annotation information, training the model through a supervised learning algorithm, determining the model parameters, and obtaining a multidimensional growth parameter model.

[0027] Furthermore, in the data analysis module, temporal feature extraction is performed using RNN (Recurrent Neural Network), with LSTM (Long Short-Term Memory) being the preferred method. Spatial feature processing uses CNN (Convolutional Neural Network), and pattern recognition is performed using the Random Forest algorithm. Specifically, the data analysis module performs temporal feature processing on the preprocessed data using an RNN, extracts spatial features from the preprocessed data using a CNN, and performs pattern recognition on the crop feature data using the Random Forest algorithm to obtain a multidimensional growth parameter model, which is a Random Forest model.

[0028] Using RNN neural networks for temporal feature extraction of crops can effectively adapt to the time series structure of preprocessed data, thereby more accurately capturing the dynamic features of multi-source data changing over time, which is beneficial to improving the accuracy of multi-dimensional growth parameter models. Employing LSTM neural networks, due to their gating mechanism, allows for selective memorization of key information and suppression of irrelevant short-term noise, thus more effectively capturing the long-term trends and stage-specific features of preprocessed data over time, further improving the accuracy of subsequent pattern recognition. Using the random forest algorithm for pattern recognition of crop feature data can effectively improve the accuracy of classification results and fully utilize the nonlinear modeling capabilities of the random forest algorithm to more effectively integrate multi-source data. It also exhibits strong robustness to nonlinear relationships between data, improving model training efficiency and inference accuracy, thereby enhancing the accuracy and efficiency of data processing.

[0029] Multidimensional growth parameter models for crops are used to describe their growth status. For different crops or different growth stages of the same crop, multidimensional growth parameter models can be selectively configured with morphological, physiological, or yield-related models based on their phenological characteristics and specific growth management needs. For example, they can include models related to appearance and morphology, and models for yield prediction. Different crops have different multidimensional growth parameter models. For instance, the multidimensional growth parameter model for corn can include developmental models, grain weight growth models, and dry matter distribution models. The multidimensional parameter module for sugar beets can include morphological development models, physiological and biochemical models, and yield formation models. Among these, the morphological development model corresponds to the phenotypic characteristics of the stems, leaves, and roots of sugar beets, and is used to describe the laws governing stem and leaf growth and root development (such as the changes in stem and leaf length, number of leaves, and root length). This can include growth stage models, root weight models, and root length models. Physiological and biochemical models correspond to intrinsic metabolic processes such as photosynthesis, water, and nutrients, and are used to describe parameters such as photosynthetic rate, water use efficiency, and nutrient absorption dynamics. These models can include plant water physiology models (such as water absorption dynamics models), growth rate models, and plant nutrient requirement models. Yield formation models correspond to the final output and are used to describe indicators such as yield per unit area, sugar content per kilogram, and root mass. These models can include yield prediction models and sugar content models.

[0030] The multidimensional growth parameter model for characterizing crop growth status, constructed based on crop characteristic data, can comprehensively cover the growth status of crops. Furthermore, the model is constructed for specific characteristic data of crops, which helps to improve the accuracy and reliability of crop growth prediction.

[0031] See also Figure 1 It is known that the system includes a digital twin engine layer 130, which is used to construct a digital twin model of crop growth based on multidimensional growth parameter models, historical air-space-ground multi-source data of crops, and regional climate characteristic information, and to predict the growth status of crops based on the digital twin model of crop growth.

[0032] Specifically, the digital twin engine layer may include: a model building module, a state prediction module, an optimization module, and a feedback calibration module.

[0033] The model building module is used to construct a digital twin model of crop growth based on a multidimensional growth parameter model, historical air-space-ground multi-source data of crops, and regional climate characteristic information. Regional climate characteristic information can be obtained by analyzing historical observation data from meteorological stations or by using existing regional climate models. The crop growth digital twin model corresponds to the multidimensional parameter model. For example, if the multidimensional growth parameter model includes a morphological development model, a physiological and biochemical model, and a yield formation model, then the crop growth digital twin model consists of a morphological development twin model, a physiological and biochemical twin model, and a yield formation twin model.

[0034] The state prediction module is used to predict the growth status of crops using a digital twin model of crop growth, and obtain the prediction results. Specifically, different planting parameters can be input into the digital twin model of crop growth, and the model can predict the growth status of crops to obtain the corresponding prediction results.

[0035] The optimization module is used to adjust the planting parameters of virtual crops in the digital twin model of crop growth based on prediction results and crop production management needs. These planting parameters include irrigation intensity and fertilization frequency. Crop production management needs can include irrigation requirements, fertilization requirements, the determination of suitable sowing dates, and the determination of harvest dates, among others.

[0036] The feedback calibration module is used to update the model parameters of the crop growth digital twin model using real-time multi-source data from air, space, and ground.

[0037] In this embodiment, the digital twin engine layer, through the configuration of the above-mentioned multiple modules, can continuously adjust and optimize the growth state of virtual crops based on machine learning and crop growth mechanisms, so as to keep it highly consistent with the growth state of actual crops, thereby effectively improving the model accuracy and thus improving the accuracy of the crop growth prediction system.

[0038] See also Figure 1 It is known that the system also includes an intelligent decision layer 140, which is used to simulate the growth process and response of crops under different planting strategies based on the digital twin model of crop growth, obtain simulation results, and generate planting decision suggestions based on the simulation results.

[0039] Planting strategies can be determined based on user needs. These strategies include irrigation, fertilization, and sowing plans for different weather conditions. Specifically, irrigation, fertilization, and sowing strategies differ under normal weather conditions and under extreme weather conditions. For example, during periods of continuous rainfall, the system will automatically adjust the irrigation strategy (reducing irrigation frequency or volume) to minimize the harmful effects of excessive water on crops. One planting strategy can be matched with multiple different planting parameters.

[0040] In implementation, a set of planting parameters matching the current planting strategy can be input into a digital twin model of crop growth to obtain the corresponding simulation results. If the simulation results match the target planting outcome, the planting parameters are maintained. If the simulation results do not match the target planting outcome, different planting parameters are adjusted for the same planting strategy through an intelligent decision-making layer, or the planting strategy is adjusted, and the planting parameters obtained from each adjustment are input into the digital twin model of crop growth again to obtain the corresponding simulation results after adjustment. After multiple (e.g., tens of thousands) adjustments of planting parameters until the target planting outcome is achieved, the corresponding optimal planting parameters are determined.

[0041] Planting decision-making recommendations can include irrigation recommendations, fertilization recommendations, and harvesting recommendations. Irrigation recommendations can be generated based on plant water physiology twin models and growth rate twin models within digital twin models; fertilization recommendations can be generated based on plant nutrient requirement twin models and growth stage twin models within digital twin models; and harvesting recommendations can be generated based on plant growth cycles and market demand forecasts. The correspondence between different types of planting decision-making recommendations and various models is shown in Table 1 below.

[0042]

[0043] Table 1. Correspondence between decision types and different models The hypothesis-verification experiment function provided by the intelligent decision-making layer can simulate the response results of crops under different planting strategies in a virtual environment, and determine the optimal planting parameters that meet the preset goals. This provides more accurate and scientific decision-making suggestions for crop growth management, effectively improves the efficiency of crop production management, and greatly reduces decision-making costs and agricultural production management costs.

[0044] See also Figure 1 It is known that the system also includes a visualization layer 150, which is used to visualize the virtual growth status of crops based on the simulation results.

[0045] Furthermore, the visualization layer displays content including: a panoramic view of the field, 3D models of individual crops, growth curves, and spatiotemporal distribution maps of crop growth indicators. The panoramic view can be generated using 3D scene reconstruction algorithms; the 3D models of individual crops can be constructed based on crop growth mechanism models; the growth curves can use line graphs or bar charts to represent the growth rates of different plant parts; and the spatiotemporal distribution maps can use heatmaps or bubble charts to represent the spatiotemporal changes of crop growth indicators.

[0046] Example 2 This specification provides a sugar beet growth prediction system based on digital twins, comprising: a data acquisition layer, a data processing layer, a digital twin engine layer, an intelligent decision-making layer, and a visualization layer.

[0047] The system comprises the following layers: a data acquisition layer for collecting multi-source data on sugar beets from air, space, and ground; a data processing layer for extracting features from the multi-source data to obtain sugar beet characteristic data, and constructing a multi-dimensional growth parameter model representing the growth status of sugar beets based on this data. This multi-dimensional growth parameter model includes a morphological development model, a physiological and biochemical model, and a yield formation model; a digital twin engine layer for constructing a digital twin model of sugar beet growth based on the multi-dimensional growth parameter model, historical multi-source data on sugar beets from air, space, and ground, and regional climate characteristics, and predicting the growth status of sugar beets based on this digital twin model; an intelligent decision-making layer for simulating the growth process and response of sugar beets under different planting strategies based on the digital twin model, obtaining simulation results, and generating planting decision suggestions based on these results. Planting strategies include irrigation strategies, fertilization plans, and sowing plans for different weather conditions; and a visualization layer for visually displaying the virtual growth status of sugar beets based on the simulation results.

[0048] This embodiment uses a beet growth prediction system based on digital twins and Figure 1 The structure and principle of the crop growth prediction system based on digital twins shown are the same, and will not be described again here.

[0049] Example 3 like Figure 2 As shown in the embodiments of this specification, a method for predicting crop growth based on digital twins is provided. The execution subject of this method can be a terminal device or a server. The terminal device can be a mobile terminal device such as a mobile phone or tablet computer, or a computer device such as a laptop or desktop computer. The server can be a single server or a server cluster composed of multiple servers. The method specifically includes the following steps: In step S1, multi-source data of crops from air, space, and ground are acquired; In step S2, feature extraction is performed based on multi-source data from air, space, and ground to obtain crop feature data, and a multi-dimensional growth parameter model representing the growth status of crops is constructed based on the crop feature data. In step S3, a digital twin model of crop growth is constructed based on a multidimensional growth parameter model, historical air-space-ground multi-source data of crops, and regional climate characteristic information, and the growth status of crops is predicted based on the digital twin model of crop growth. In step S4, based on the digital twin model of crop growth, the growth process and response of crops under different planting strategies are simulated to obtain simulation results. Planting decision suggestions are generated based on the simulation results. The planting strategies include irrigation strategies, fertilization plans and sowing plans under different weather conditions. In step S5, the virtual growth status of crops is visualized based on the simulation results.

[0050] Furthermore, the processing in step S4 above can take many forms. The following is one optional processing method, which can be found in the following steps S41-S44: In step S41, one or more sets of planting parameters matching the current planting strategy are determined, and the determined current planting parameters are input into the crop growth digital twin model to obtain the simulation results corresponding to the current planting parameters. In step S42, if the simulation results corresponding to the current planting parameters do not meet the target planting results, the planting parameters and / or planting strategy are adjusted to obtain the adjusted planting parameters, and the adjusted planting parameters are input into the crop growth digital twin model to obtain the simulation results corresponding to the adjusted planting parameters. In step S43, the planting parameters and / or planting strategy are repeatedly adjusted until the simulation results obtained meet the target planting results, and the planting parameters corresponding to the target planting results are determined. In step S44, planting decision suggestions are generated based on the planting parameters corresponding to the target planting results.

[0051] Furthermore, step S4 above also includes step S45: if the simulation result corresponding to the current planting parameters matches the target planting result, then the current planting parameters are maintained.

[0052] Furthermore, the processing in step S1 above can take many forms. The following provides an optional processing method, which can be found in the following steps S11-S14: In step S11, remote multispectral image data of crops are acquired based on multispectral cameras mounted on satellites and drones; In step S12, environmental parameters of the crop are collected, including soil moisture, temperature, light intensity and CO2 concentration. In step S13, biochemical indicators of crops are collected, including chlorophyll content and protein content. In step S14, the system receives multi-source data from the multispectral image acquisition module, the microenvironment parameter acquisition module, and the biochemical index acquisition module via the aggregation node, the smart gateway, and the cloud server, and transmits the multi-source data to the data processing layer.

[0053] It should be noted that the execution order of the above steps S11-S13 can be executed in parallel or the order can be changed. This embodiment does not limit this.

[0054] Furthermore, the multidimensional growth parameter model in step S2 includes: a morphological development model, a physiological and biochemical model, and a yield formation model.

[0055] Furthermore, the processing in step S2 above can take many forms. The following is one optional processing method, which can be found in the following steps S21-S22: In step S21, the multi-source data from air, space, and ground is preprocessed to obtain preprocessed data. The preprocessing includes one or more of the following: standardization, normalization, and missing value handling. In step S22, feature extraction processing is performed on the preprocessed data to obtain crop feature data, and pattern recognition processing is performed on the crop feature data to obtain a multidimensional growth parameter model representing the crop growth status. The feature extraction processing includes: temporal feature extraction processing and spatial feature extraction processing.

[0056] Furthermore, in step S22, the temporal feature extraction process uses an RNN neural network, the spatial feature processing uses a CNN neural network, and the pattern recognition process uses a random forest algorithm.

[0057] Furthermore, the processing in step S3 above can take many forms. The following is one optional processing method, which can be found in the following steps S31-S34: In step S31, a digital twin model of crop growth is constructed based on a multidimensional growth parameter model, historical air-space-ground multi-source data of crops, and regional climate characteristic information. In step S32, the growth status of crops is predicted using a digital twin model of crop growth, and the prediction results are obtained. In step S33, the planting parameters of the virtual crops in the crop growth digital twin model are adjusted according to the prediction results and crop production management needs. The planting parameters include irrigation intensity and fertilization frequency. In step S34, the model parameters of the crop growth digital twin model are updated using real-time multi-source data from air, space, and ground of the crop.

[0058] For the parts not described in detail in this embodiment, please refer to Embodiment 1 and Embodiment 2. The three embodiments can be referred to each other, and will not be repeated here.

[0059] Example 4 The above describes the crop growth prediction system and method based on digital twins provided in the embodiments of this specification. Based on the same idea, the embodiments of this specification also provide an electronic device, such as... Figure 3 As shown. Electronic devices can vary considerably due to differences in configuration or performance, and may include one or more processors 301 and memory 302. Memory 302 may store one or more application programs or data. Memory 302 may be temporary or persistent storage. The application programs stored in memory 302 may include one or more modules (not shown), each module may include a series of computer-executable instructions for the electronic device. Furthermore, processor 301 may be configured to communicate with memory 302 and execute the series of computer-executable instructions in memory 302 on the electronic device. The electronic device may also include one or more power supplies 303, one or more wired or wireless network interfaces 304, one or more input / output interfaces 305, and one or more keyboards 306.

[0060] Specifically, in this embodiment, the electronic device includes a memory and one or more programs, wherein one or more programs are stored in the memory, and one or more programs may include one or more modules, and each module may include a series of computer-executable instructions for use in the electronic device, and is configured to be executed by one or more processors. The one or more programs include computer-executable instructions for performing the following: Acquire multi-source data on crops from air, space, and ground; Feature extraction and processing are performed on multi-source data from air, space, and ground to obtain crop feature data, and a multi-dimensional growth parameter model representing the growth status of crops is constructed based on the crop feature data. Based on multidimensional growth parameter models, historical air-space-ground multi-source data of crops, and regional climate characteristic information, a digital twin model of crop growth is constructed, and the growth status of crops is predicted based on the digital twin model of crop growth. Based on a digital twin model of crop growth, the growth process and response of crops under different planting strategies are simulated to obtain simulation results. Based on the simulation results, planting decision suggestions are generated, including irrigation strategies, fertilization plans and sowing plans under different weather conditions. The virtual growth status of crops is visualized based on the simulation results.

[0061] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A crop growth prediction system based on digital twins, characterized in that, The system includes: The data acquisition layer is used to collect multi-source data on crops from air, space, and ground. The data processing layer is used to perform feature extraction processing based on the multi-source data from air, space, and ground to obtain crop feature data, and to construct a multi-dimensional growth parameter model representing the growth status of crops based on the crop feature data. The digital twin engine layer is used to construct a digital twin model of crop growth based on the multidimensional growth parameter model, historical air-space-ground multi-source data of crops, and regional climate characteristic information, and to predict the growth status of crops based on the digital twin model of crop growth. The intelligent decision-making layer is used to simulate the growth process and response of crops under different planting strategies based on the crop growth digital twin model, obtain simulation results, and generate planting decision suggestions based on the simulation results. The planting strategies include irrigation strategies, fertilization schemes, and sowing schemes under different weather conditions. The visualization layer is used to visualize the virtual growth status of crops based on the simulation results.

2. The crop growth prediction system according to claim 1, characterized in that, The multidimensional growth parameter model includes: morphological development model, physiological and biochemical model, and yield formation model.

3. The crop growth prediction system according to claim 1, characterized in that, The data acquisition layer includes: The multispectral image acquisition module is used to acquire remote multispectral image data of crops based on multispectral cameras mounted on satellites and drones; The microenvironment parameter acquisition module is used to collect environmental parameters of crops, including soil moisture, temperature, light intensity, and CO2 concentration. A biochemical indicator acquisition module is used to collect biochemical indicators of crops, including chlorophyll content and protein content. The integrated air-space-ground sensor network is used to receive multi-source air-space-ground data from the multispectral image acquisition module, microenvironment parameter acquisition module, and biochemical index acquisition module through aggregation nodes, smart gateways, and cloud servers, and transmit the multi-source air-space-ground data to the data processing layer.

4. The crop growth prediction system according to claim 1, characterized in that, The data processing layer includes: The data preprocessing module is used to preprocess the multi-source data from air, space, and ground to obtain preprocessed data. The preprocessing includes one or more of the following: standardization, normalization, and missing value processing. The data analysis module is used to perform feature extraction processing on the preprocessed data to obtain crop feature data, and to perform pattern recognition processing on the crop feature data to obtain a multidimensional growth parameter model characterizing the crop growth status. The feature extraction processing includes: temporal feature extraction processing and spatial feature extraction processing.

5. The crop growth prediction system according to claim 4, characterized in that, In the data analysis module, the temporal feature extraction process uses an RNN neural network, the spatial feature processing uses a CNN neural network, and the pattern recognition process uses a random forest algorithm.

6. The crop growth prediction system according to claim 1, characterized in that, The digital twin engine layer includes: The model building module is used to construct a digital twin model of crop growth based on the multidimensional growth parameter model, historical air-space-ground multi-source data of crops, and regional climate characteristic information. The state prediction module is used to predict the growth state of crops using the crop growth digital twin model and obtain the prediction results. The optimization module is used to adjust the planting parameters of the virtual crops in the crop growth digital twin model according to the prediction results and crop production management needs. The planting parameters include irrigation intensity and fertilization frequency. The feedback calibration module is used to update the model parameters of the crop growth digital twin model using real-time multi-source data from air, space, and ground.

7. The crop growth prediction system according to claim 1, characterized in that, The visualization display layer displays one or more of the following: a panoramic view of the field, a three-dimensional model of an individual crop, a growth curve, and a spatiotemporal distribution map of crop growth indicators.

8. A crop growth prediction method based on digital twins, characterized in that, The method includes: Acquire multi-source data on crops from air, space, and ground; Based on the multi-source data from air, space, and ground, feature extraction processing is performed to obtain crop feature data, and a multi-dimensional growth parameter model characterizing the growth status of crops is constructed based on the crop feature data. Based on the multidimensional growth parameter model, historical air-space-ground multi-source data of crops, and regional climate characteristic information, a digital twin model of crop growth is constructed, and the growth status of crops is predicted based on the digital twin model of crop growth. Based on the aforementioned digital twin model of crop growth, the growth process and response of crops under different planting strategies are simulated to obtain simulation results. Based on the simulation results, planting decision suggestions are generated. The planting strategies include irrigation strategies, fertilization schemes, and sowing schemes under different weather conditions. The virtual growth status of crops is visualized based on the simulation results.

9. The crop growth prediction method according to claim 8, characterized in that, Based on the aforementioned digital twin model of crop growth, the growth process and response of crops under different planting strategies are simulated to obtain simulation results. Based on these simulation results, planting decision recommendations are generated, including: Determine one or more sets of planting parameters that match the current planting strategy, input the determined current planting parameters into the crop growth digital twin model, and obtain the simulation results corresponding to the current planting parameters; If the simulation results corresponding to the current planting parameters do not meet the target planting results, the planting parameters and / or planting strategy are adjusted to obtain the adjusted planting parameters, and the adjusted planting parameters are input into the crop growth digital twin model to obtain the simulation results corresponding to the adjusted planting parameters. Repeatedly adjust the planting parameters and / or planting strategy until the simulation results obtained meet the target planting results, and determine the planting parameters corresponding to the target planting results; Planting decision recommendations are generated based on the planting parameters corresponding to the target planting results.

10. An electronic device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the crop growth prediction method based on digital twins as described in claim 8 or 9.

Images