Agricultural environment information acquisition control method, system, equipment and medium
By collecting and analyzing farmland remote sensing images and environmental data, identifying crop types and growth status, setting personalized thresholds and generating control strategies, the problem that the existing technology cannot manage different monitoring areas in a refined manner, and improving farmland management efficiency and accuracy Optimize crop growth and resource use.
Patent Information
- Application Number
- CN202411077990.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-07
- Publication Date
- 2024-11-26
AI Technical Summary
The existing agricultural automation control system cannot provide refined monitoring and management strategies for different monitoring areas, resulting in reduced farmland management efficiency and accuracy.
By collecting remote sensing image information and environmental data from multiple monitoring areas in the farmland, using machine learning algorithms to identify crop types and growth status, setting a personalized environmental abnormality threshold range, and generating corresponding control strategies to communicate to the intelligent control system for execution .
It has achieved refined management and efficient monitoring of the farmland environment, improved the overall efficiency and accuracy of farmland management, optimized crop growth conditions, reduced resource consumption, and improved the intelligence and sustainability of agricultural production.
Smart Images

Figure CN119027262A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of agricultural automatic control technology, and in particular to an agricultural environmental information collection and control method, system, equipment and medium. Background Art
[0002] With the growth of the global population and the increasing concern about food security, the development of precision agriculture technology has become particularly important. The core concept of precision agriculture is to improve crop yield and quality while reducing resource consumption through refined management of farmland. In this process, farmland environment monitoring and control technology plays a vital role.
[0003] At present, in order to manage large farmland more effectively, modern agricultural practices tend to divide farmland into multiple monitoring areas, each of which grows different types of crops or different growth stages of the same crop, so the requirements for environmental conditions will also be different. This zoning management method has been widely recognized and not only helps to achieve more accurate resource allocation and environmental control, but also improves overall production efficiency and economic benefits.
[0004] Regarding the above-mentioned related technologies, the applicant found that due to the differences in environmental conditions, crop types and growth conditions in different monitoring areas, common agricultural automation control systems are often unable to provide refined monitoring and management strategies for each monitoring area, reducing the efficiency and accuracy of farmland management. Summary of the invention
[0005] In order to improve the efficiency and accuracy of farmland management, the present application provides an agricultural environment information collection and control method, system, equipment and medium.
[0006] In the first aspect, the present application provides an agricultural environmental information collection and control method, which adopts the following technical solution: A method for collecting and controlling agricultural environmental information, the method comprising: Collect remote sensing image information and environmental data information of multiple monitoring areas in farmland and perform data preprocessing; Extracting features based on the remote sensing image information to obtain crop features in each monitoring area; Based on a machine learning algorithm, identifying the crop type and crop growth status corresponding to the crop characteristics; Based on a preset mapping table, determining an environmental anomaly threshold range for each monitoring area according to the crop type and crop growth status; According to the environmental anomaly threshold range of each monitoring area, the environmental data information of each monitoring area is analyzed for anomalies, and the environmental analysis results of each monitoring area are obtained; In response to the environmental analysis result of the monitoring area being an abnormal environmental result, generating a control strategy according to the abnormal environmental result; The control strategy is sent to the intelligent control system corresponding to the monitoring area.
[0007] By adopting the above technical solution, remote sensing images and environmental data of multiple monitoring areas in the farmland are collected and analyzed in real time, and the crop type and growth status of each area can be automatically identified, and personalized environmental abnormality thresholds can be set accordingly. Through abnormal analysis of environmental data, the system can generate corresponding control strategies for the specific conditions of each monitoring area and transmit them to the corresponding intelligent control system for execution, thereby realizing refined management and efficient monitoring of the farmland environment and improving the overall efficiency and accuracy of farmland management.
[0008] Optionally, the step of extracting features according to the remote sensing image information to obtain crop features of each monitoring area includes: Segmenting the image into a plurality of super-pixel regions according to the remote sensing image information, wherein pixels in each super-pixel region have similar features; Extract multispectral features from each superpixel region; Calculate the texture features in each superpixel region based on the gray level co-occurrence matrix; Extract morphological features of each superpixel region; Combining the multispectral features, texture features and morphological features into a feature vector; A key feature subset is selected from the feature vector as crop features based on a feature selection method.
[0009] By adopting the above technical solutions, it is possible to efficiently and accurately extract crop features from remote sensing images, so as to accurately distinguish different types of crops and evaluate the growth status of crops, providing strong support for farmland management and decision-making.
[0010] Optionally, based on a machine learning algorithm, the step of identifying the crop type and crop growth status corresponding to the crop feature includes: The crop features are input into a pre-trained crop recognition model to obtain the crop type and crop growth status corresponding to the crop features; wherein the crop recognition model is obtained through iterative training based on multiple sample images and pre-labeled crop type labels and crop growth status labels.
[0011] Optionally, the step of performing anomaly analysis on the environmental data information of each monitoring area according to the environmental anomaly threshold range of each monitoring area includes: Determine whether the environmental data information of each monitoring area exceeds the corresponding environmental abnormality threshold range, and if so, determine that the environmental data information of the monitoring area is abnormal; Determine whether the duration of abnormality in the environmental data information of the monitoring area exceeds a preset duration threshold; If it exceeds, the environmental analysis result of the monitoring area is output as an environmental abnormality result, and the corresponding abnormality type and abnormality degree are determined according to the abnormal environmental data information; If it does not exceed, the environmental analysis result of the monitoring area is output as a normal environmental result.
[0012] By adopting the above technical solutions and comprehensively applying the abnormal detection and duration judgment of environmental data, refined monitoring and analysis of farmland environmental information is achieved. By accurately judging whether the environmental data exceeds the abnormal threshold range and whether the duration of the abnormality exceeds the preset duration threshold, the system can promptly identify and respond to environmental abnormalities, thereby effectively guiding farmland management decisions. This precise monitoring and analysis helps optimize crop growth conditions, improve crop yields and quality, while reducing unnecessary resource consumption, and realizing intelligent and sustainable farmland management.
[0013] Optionally, the step of generating a control strategy according to the abnormal environment result includes: Determine the corresponding abnormality type and abnormality degree according to the environmental abnormality result; Based on the preset rule relationship, determine the corresponding environmental parameter adjustment type and environmental parameter adjustment index according to the abnormality type and abnormality degree; According to the environmental parameter adjustment type and the environmental parameter adjustment index, the target environmental parameters of the monitoring area are adjusted and a corresponding control strategy is generated.
[0014] By adopting the above technical solution, environmental parameters are automatically adjusted based on preset rule relationships, achieving effective control of the agricultural environment. This method can not only respond to various environmental anomalies in a timely manner, but also take appropriate countermeasures according to the specific circumstances of the anomalies, thereby effectively improving the crop growth environment and improving the growth quality and yield of crops. At the same time, by reducing the need for manual intervention, management costs are reduced, and the level of intelligence and sustainability of agricultural production are improved.
[0015] Optionally, the method further comprises a step of training the crop recognition model, wherein the training step comprises: Acquire a sample image data set and perform image preprocessing; wherein the sample image data set includes a plurality of sample images and pre-labeled crop type labels and crop growth status labels; Perform feature extraction on the sample image data set to obtain sample crop features; Divide sample crop features into training set, validation set and test set; Inputting the training set into a pre-built convolutional neural network for initial training; wherein the convolutional neural network includes a convolutional layer, a pooling layer, a fully connected layer and an output layer, and the output layer is used to output the predicted probability of each crop type and crop growth state; The initially trained convolutional neural network is verified and model parameters are optimized based on the verification set and the test set to obtain the trained crop recognition model.
[0016] By adopting the above technical solutions, a highly automated crop recognition model was built, which can effectively identify crop types and evaluate crop growth status from remote sensing images, providing strong support for farmland management and decision-making.
[0017] Optionally, the training set is input into a pre-built convolutional neural network for initial training, and the steps of verifying and optimizing model parameters of the initially trained convolutional neural network based on the verification set and the test set include: Inputting the training set into a pre-built convolutional neural network for iterative training, optimizing the model parameters and calculating the cross entropy loss function of the model until the cross entropy loss function meets the preset conditions, thereby obtaining a trained convolutional neural network; Validating the convolutional neural network based on the validation set, evaluating the performance of the convolutional neural network and adjusting the hyperparameters of the model according to the validation results; The prediction ability of the adjusted convolutional neural network is tested based on the test set.
[0018] By adopting the above technical solutions, the recognition accuracy of the model can be improved, so as to effectively identify crop types and evaluate crop growth status from remote sensing images, providing strong support for farmland management and decision-making.
[0019] In the second aspect, the present application provides an agricultural environmental information collection and control system, which adopts the following technical solution: An agricultural environmental information collection and control system, the system comprising: Data acquisition module, used to collect remote sensing image information and environmental data information of multiple monitoring areas in the farmland; A preprocessing module is used to preprocess remote sensing image information and environmental data information; A feature extraction module, used to extract features based on the remote sensing image information to obtain crop features in each monitoring area; An identification module, used for identifying the crop type and crop growth status corresponding to the crop characteristics based on a machine learning algorithm; An abnormal threshold range determination module is used to determine the environmental abnormality threshold range of each monitoring area according to the crop type and crop growth status based on a preset mapping table; The abnormality analysis module is used to perform abnormality analysis on the environmental data information of each monitoring area according to the environmental abnormality threshold range of each monitoring area, and obtain the environmental analysis results of each monitoring area; A control strategy generating module, configured to generate a control strategy according to the abnormal environment result in response to the environmental analysis result of the monitoring area being an abnormal environment result; The control strategy sending module is used to send the control strategy to the intelligent control system corresponding to the monitoring area.
[0020] In a third aspect, the present application provides a computer device, which adopts the following technical solution: A computer device comprises a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method as described in the first aspect.
[0021] In a fourth aspect, the present application provides a computer-readable storage medium, which adopts the following technical solution: A computer-readable storage medium stores a computer program that can be loaded by a processor and execute any one of the methods in the first aspect.
[0022] In summary, the present application includes at least one of the following beneficial technical effects: real-time collection and analysis of remote sensing images and environmental data of multiple monitoring areas in farmland, which can automatically identify the crop type and growth status of each area, and set personalized environmental abnormality thresholds accordingly. Through abnormal analysis of environmental data, the system can generate corresponding control strategies for the specific conditions of each monitoring area, and transmit them to the corresponding intelligent control system for execution, thereby realizing refined management and efficient monitoring of the farmland environment and improving the overall efficiency and accuracy of farmland management. BRIEF DESCRIPTION OF THE DRAWINGS
[0023] Figure 1 This is a first flow chart of an agricultural environment information collection and control method according to one of the embodiments of the present application.
[0024] Figure 2 This is a second flow chart of an agricultural environment information collection and control method according to one of the embodiments of the present application.
[0025] Figure 3 It is a third flow chart of an agricultural environment information collection and control method in one of the embodiments of the present application.
[0026] Figure 4 It is a fourth flow chart of an agricultural environment information collection and control method according to one of the embodiments of the present application.
[0027] Figure 5It is the fifth flow chart of an agricultural environment information collection and control method in one of the embodiments of the present application.
[0028] Figure 6 It is the sixth flow chart of an agricultural environment information collection and control method in one of the embodiments of the present application. DETAILED DESCRIPTION
[0029] In order to make the purpose, technical solutions and advantages of this application more clear, the following Figure 1-6 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.
[0030] The embodiment of the present application discloses a method for collecting and controlling agricultural environmental information.
[0031] Reference Figure 1 , a method for collecting and controlling agricultural environmental information, the method comprising: Step S101, collecting remote sensing image information and environmental data information of multiple monitoring areas in the farmland and performing data preprocessing; Among them, drones, satellites or other remote sensing equipment can be used to collect remote sensing image information in farmland. These images can be used to identify crop types and monitor crop growth conditions; environmental data such as temperature, humidity, light intensity, soil moisture, etc. can be collected through a series of ground sensors (such as soil moisture sensors, temperature and humidity sensors, light intensity sensors, etc.).
[0032] Specifically, the data preprocessing step is to clean and standardize the data, including denoising, removing outliers, filling missing values, unifying data formats, etc., to ensure data quality and the accuracy and reliability of subsequent analysis. For remote sensing image data, image enhancement operations can also be performed to improve image quality.
[0033] Step S102, extracting features based on remote sensing image information to obtain crop features in each monitoring area; In some embodiments, the extracted crop features may include multispectral features, texture features, and morphological features, etc. Multispectral features include reflectance of near-infrared band and red band, texture features include contrast and homogeneity, and morphological features include circularity and area.
[0034] Step S103, identifying the crop type and crop growth status corresponding to the crop characteristics based on a machine learning algorithm; In some embodiments, a pre-trained classification model may be used to identify the crop type based on the extracted features, and the growth status of the crop may be evaluated based on the morphological characteristics of the crop (such as leaf color, plant height, etc.); For example, if the crop characteristics of the monitored area have high near-infrared band reflectivity, moderate contrast, and high circularity, then this area is likely to be a soybean crop; if the near-infrared band reflectivity is low, the contrast is high, and the circularity is low, then this area is likely to be a corn crop.
[0035] Among them, the growth status of crops refers to the growth stage. Taking wheat as an example, it usually includes the jointing stage, heading stage, flowering stage, filling stage, maturity stage, etc.; by determining the growth status of crops, the growth condition of wheat can be evaluated more accurately, and corresponding management measures can be taken accordingly; for example, in the jointing stage and heading stage, the irrigation and fertilization management of wheat can be strengthened to ensure adequate water and nutrient supply; in the flowering stage, measures can be taken to protect crops from pests and diseases; in the filling stage, the development of grains can be monitored to ensure that the grains are fully filled; in the maturity stage, they can be harvested in time to obtain the best yield and quality.
[0036] Step S104, determining the environmental abnormality threshold range of each monitoring area according to the crop type and crop growth status based on the preset mapping table; Among them, a mapping relationship table between crop types, growth stages and environmental parameters can be pre-established in combination with historical experience and data analysis results, and different threshold ranges can be set according to crop characteristics and growth stages, such as soil moisture threshold range, light intensity threshold range, etc. For example, for wheat, the threshold range of soil moisture may be set to 15%~25% during the jointing stage, and to 20%~30% during the grain filling stage.
[0037] Understandably, more precise environmental management can be achieved based on the specific needs of different crop types, while adjusting the threshold range in conjunction with changes in crop growth stages improves adaptability; Step S105, performing an abnormality analysis on the environmental data information of each monitoring area according to the environmental abnormality threshold range of each monitoring area, and obtaining an environmental analysis result of each monitoring area; Specifically, the environmental data of each monitoring area is compared with the corresponding threshold range to identify abnormal data. By performing statistical analysis on the abnormal data, the type and severity of the abnormality are identified to obtain the environmental analysis result.
[0038] Step S106, in response to the environmental analysis result of the monitoring area being an abnormal environmental result, generating a control strategy according to the abnormal environmental result; Step S107, sending the control strategy to the intelligent control system corresponding to the monitoring area.
[0039] In some embodiments, according to the abnormality type (such as temperature and humidity abnormalities, soil moisture abnormalities, light intensity abnormalities, etc.) and the degree of abnormality (the degree exceeding the environmental abnormality threshold range) in the environmental abnormality results, corresponding control measures can be selected from a preset strategy library, and corresponding control strategies can be generated and executed through an intelligent control system, such as adjusting the working mode of the irrigation system, controlling the shading system, ventilation and cooling system and other equipment, to achieve temperature and humidity regulation.
[0040] Exemplarily, if the soil moisture is below a threshold range, a control strategy for increasing the irrigation amount is generated, and the intelligent control system adjusts the configuration of the automated irrigation system and increases the amount of water to the monitored area.
[0041] In the above implementation, remote sensing images and environmental data of multiple monitoring areas in the farmland are collected and analyzed in real time, and the crop type and growth status of each area can be automatically identified, and personalized environmental abnormality thresholds can be set accordingly. Through abnormal analysis of environmental data, the system can generate corresponding control strategies for the specific conditions of each monitoring area, and transmit them to the corresponding intelligent control system for execution, thereby realizing refined management and efficient monitoring of the farmland environment and improving the overall efficiency and accuracy of farmland management.
[0042] Reference Figure 2 As an implementation of step S102, the step of extracting features based on remote sensing image information to obtain crop features in each monitoring area includes: Step S201, dividing the image into a plurality of super-pixel regions according to remote sensing image information, wherein pixels in each super-pixel region have similar features; In some embodiments, a superpixel segmentation algorithm (such as the SLIC algorithm) can be used to segment the remote sensing image into multiple superpixel regions; the SLIC algorithm is a fast image segmentation method that can segment the image into regions with similar colors and spatial positions. These regions are called superpixels. The size and shape of the superpixels can be adjusted as needed to adapt to different types of crops and monitoring needs.
[0043] It can be understood that through superpixel segmentation, the image can be simplified into a group of regions with similar features, which is helpful for subsequent feature extraction and crop type identification, and can reduce the amount of calculation and improve the processing speed.
[0044] Step S202, extracting multispectral features from each superpixel region; Among them, the extracted multispectral features include but are not limited to the reflectance of red, green, blue (RGB) and near-infrared (NIR) bands. Multispectral features reflect the spectral changes of crops and are an important basis for crop type identification and growth status assessment. Reflectance features can be used to assess the growth status of crops, such as chlorophyll content, health status, etc.
[0045] Step S203, calculating the texture features in each superpixel region based on the gray level co-occurrence matrix; Specifically, the Gray Level Co-occurrence Matrix (GLCM) can be used to calculate the texture features within each superpixel area; GLCM is a commonly used texture analysis method that calculates texture features such as contrast, homogeneity, energy, etc. by analyzing the probability of occurrence of adjacent pixels in grayscale.
[0046] Among them, since crops of different types and growth states usually have different texture features, texture features reflect the local structural information of the image and help to identify the surface texture of the crop, which is very important for the identification of crop types.
[0047] Step S204, extracting morphological features of each superpixel region; Among them, morphological features such as area, perimeter, circularity, etc. Morphological features describe the geometry of the superpixel region and help identify the shape and size of crops. For example, circularity can be used to distinguish between circular crops and non-circular crops.
[0048] Step S205, combining the multispectral features, texture features and morphological features into a feature vector; Among them, the fused feature vector contains the key information of each superpixel area, which is used for subsequent crop type identification and growth status assessment.
[0049] Step S206: selecting a key feature subset from the feature vector as crop features based on a feature selection method.
[0050] In some embodiments, a feature selection method (such as recursive feature elimination (RFE)) may be used to select a subset of key features from the feature vector as crop features, and the feature selection method aims to identify the features that contribute most to crop type identification. For example, the feature selection method can find that the reflectivity, contrast, and circularity in the near infrared band contribute most to crop type identification.
[0051] For example, the reflectivity of the near-infrared band can be used to distinguish different crop types, because different crops have different chlorophyll contents, resulting in different reflectivity in the near-infrared band; contrast can be used to identify texture features on the crop surface, which is also very useful for distinguishing crop types; circularity can be used to distinguish between round crops (such as soybeans) and non-round crops (such as corn).
[0052] It is understandable that feature selection can reduce the number of redundant features, improve the efficiency and accuracy of the model, and the extracted key feature subset can improve the interpretability and generalization ability of the model.
[0053] In the above implementation, crop features are efficiently and accurately extracted from remote sensing images, so as to accurately distinguish different types of crops and evaluate the growth status of crops, providing strong support for farmland management and decision-making.
[0054] As an implementation of step S103, based on a machine learning algorithm, the step of identifying the crop type and crop growth status corresponding to the crop characteristics includes: The crop features are input into a pre-trained crop recognition model to obtain the crop type and crop growth status corresponding to the crop features; wherein the crop recognition model is iteratively trained based on multiple sample images and pre-labeled crop type labels and crop growth status labels.
[0055] Reference Figure 3 As an implementation of the crop recognition model, the training steps of the crop recognition model include: Step S301, obtaining a sample image data set and performing image preprocessing; wherein the sample image data set includes a plurality of sample images and pre-labeled crop type labels and crop growth status labels; Specifically, before model training, a large number of high-quality remote sensing images containing different crop types can be collected from farmland, and necessary preprocessing of the images can be performed, such as resizing, brightness adjustment, color correction, etc., to ensure that the image quality is consistent and suitable as training data.
[0056] Step S302, extracting features from the sample image data set to obtain sample crop features; The feature extraction step includes using a superpixel segmentation algorithm (such as the SLIC algorithm) to segment the sample image into multiple superpixel regions, extracting multispectral features from each superpixel region, calculating the texture features in each superpixel region based on the gray level co-occurrence matrix, extracting the morphological features of each superpixel region, combining these features into a feature vector, and using a feature selection method (such as recursive feature elimination RFE) to select a key feature subset from the feature vector as the sample crop feature. For the specific feature extraction steps, refer to the feature extraction steps for remote sensing image information in the above steps S201 to S206.
[0057] Step S303, dividing the sample crop features into a training set, a validation set and a test set; The extracted sample crop features are randomly divided into three parts: training set, validation set and test set. The training set is used to train the model, the validation set is used to adjust the model parameters, and the test set is used to evaluate the final performance of the model. By dividing the data set, the effectiveness of model training can be ensured, and the generalization ability of the model can be objectively evaluated. The existence of the validation set and the test set helps prevent the model from overfitting and ensures that the model can perform well on unknown data.
[0058] Step S304, inputting the training set into a pre-built convolutional neural network for initial training; wherein the convolutional neural network includes a convolutional layer, a pooling layer, a fully connected layer and an output layer, and the output layer is used to output the predicted probability of each crop type and crop growth state; In one of the embodiments of the present application, a convolutional neural network is selected as the basis for model training. A convolutional neural network (CNN) is a deep learning model specifically designed to process data with a grid structure (such as the image data in the present application). By stacking multiple layers of convolutional layers and pooling layers, it can automatically learn multi-level features in remote sensing images of crops.
[0059] Specifically, in the convolutional neural network architecture, the convolution layer is used to capture local features of the image. Each convolution layer contains multiple filters (convolution kernels), and each filter performs a convolution operation with a part of the input image to generate a feature map. These filters can detect different features in the image, such as edges, textures, etc.; the pooling layer is used to reduce the size of the feature map while retaining the most important information. Common pooling operations include maximum pooling and average pooling, which can help reduce the amount of computation and enhance the translation invariance of the model; the fully connected layer is located at the end of the network and is used to integrate the previously extracted features for final classification. All features are mapped to a vector of a fixed length, which is used as the input of the output layer; the output layer is after the fully connected layer and is used to output the final prediction result. The activation function of the output layer can use the softmax function to calculate the probability of each category.
[0060] Step S305, verifying and optimizing the model parameters of the initially trained convolutional neural network based on the verification set and the test set to obtain a trained crop recognition model.
[0061] The validation set is used to verify the convolutional neural network after initial training, and the hyperparameters (such as learning rate, batch size, etc.) are adjusted to optimize the model performance. At the same time, the performance of the final model is evaluated on the test set, including indicators such as accuracy, recall rate, and F1 score.
[0062] In the above implementation, a highly automated crop recognition model is constructed, which can effectively identify crop types and evaluate crop growth status from remote sensing images, providing strong support for farmland management and decision-making.
[0063] Reference Figure 4 As an implementation method of step S304 and step S305, the training set is input into a pre-built convolutional neural network for initial training, and the steps of verifying the initially trained convolutional neural network and optimizing model parameters based on the verification set and the test set include: Step S401, inputting the training set into a pre-built convolutional neural network for iterative training, optimizing the model parameters and calculating the cross entropy loss function of the model until the cross entropy loss function meets the preset conditions, thereby obtaining a trained convolutional neural network; Specifically, a pre-built convolutional neural network (CNN) is iteratively trained using training set data. In each iteration, the convolutional neural network receives a batch of training samples as input and calculates prediction results through forward propagation. Based on the difference between the prediction results and the actual labels, for example, using the cross entropy loss function as a measure, the weights and biases of the network are adjusted through the back propagation algorithm to minimize the loss function, and the training is iterated until the loss function converges.
[0064] Step S402, verifying the convolutional neural network based on the verification set, evaluating the performance of the convolutional neural network and adjusting the hyperparameters of the model according to the verification results; Specifically, the trained convolutional neural network is verified using the validation set data, and the performance indicators such as the loss and accuracy of the model on the validation set are evaluated. According to the validation results, the hyperparameters of the model are adjusted, such as the learning rate, batch size, regularization coefficient, etc. In addition, an early stopping strategy can be used, that is, the training is terminated early when the performance on the validation set no longer improves.
[0065] Step S403: testing the prediction ability of the adjusted convolutional neural network based on the test set.
[0066] Among them, the various performance indicators of the model on the test set, such as accuracy, recall rate, F1 score, etc., are calculated. The test results can objectively reflect the performance of the model on unknown data, and by verifying the effectiveness and practicality of the model, ensure that the model can accurately identify the crop type and growth status in the farmland.
[0067] In the above implementation, the recognition accuracy of the model can be improved, thereby effectively identifying crop types and evaluating crop growth status from remote sensing images, providing strong support for farmland management and decision-making.
[0068] Reference Figure 5 As an implementation of step S105, the step of performing abnormal analysis on the environmental data information of each monitoring area according to the environmental abnormality threshold range of each monitoring area includes: Step S501, determine whether the environmental data information of each monitoring area exceeds the corresponding environmental abnormality threshold range, if exceeded, jump to step S502; if not exceeded, jump to step S505; Step S502, determining whether the environmental data information in the monitoring area is abnormal; Among them, for each monitoring area, real-time environmental data information is obtained, and these data are compared with the preset environmental abnormality threshold range to determine whether they exceed the normal range, so as to quickly identify abnormal conditions in the environmental data and provide a basis for subsequent abnormality analysis and response.
[0069] Step S503, determining whether the duration of abnormality in the environmental data information of the monitoring area exceeds a preset duration threshold; If it exceeds, jump to step S504; if it does not exceed, jump to step S505; Step S504, outputting the environmental analysis result of the monitoring area as an environmental abnormality result, and determining the corresponding abnormality type and abnormality degree according to the abnormal environmental data information; Among them, the time point when the environmental data exceeds the abnormal threshold range is recorded, the duration of the environmental data exceeding the abnormal threshold range is calculated, and compared with the preset duration threshold to determine whether measures need to be taken.
[0070] Specifically, when the monitored environmental data exceeds the abnormal threshold range and lasts for more than the preset time threshold, the system will automatically identify the specific abnormality type (such as high temperature, insufficient humidity, etc.) and determine the degree of abnormality based on the degree of exceeding the threshold range, such as minor abnormality, moderate abnormality or severe abnormality.
[0071] Step S505, outputting the environmental analysis result of the monitoring area as a normal environmental result.
[0072] Among them, if there is no abnormality in the environmental data information or the duration of the abnormality does not exceed the preset duration threshold, the output environmental analysis result is a normal environmental result.
[0073] It is understandable that by judging the duration of the abnormality, unnecessary warnings caused by short-term environmental fluctuations can be avoided, ensuring that control measures are only taken when the environmental abnormality lasts long enough, thereby reducing resource waste.
[0074] In the above implementation, the comprehensive application of abnormal detection of environmental data and judgment of duration realizes the refined monitoring and analysis of farmland environmental information. By accurately judging whether the environmental data exceeds the abnormal threshold range and whether the duration of the abnormality exceeds the preset duration threshold, the system can promptly identify and respond to environmental abnormalities, thereby effectively guiding farmland management decisions. This precise monitoring and analysis helps to optimize crop growth conditions, improve crop yields and quality, while reducing unnecessary resource consumption, and realize intelligent and sustainable farmland management.
[0075] Reference Figure 6 As an implementation of step S106, the step of generating a control strategy according to the abnormal environment result includes: Step S601, determining the corresponding abnormality type and abnormality degree according to the environmental abnormality result; For example, if the monitored temperature is higher than the upper limit of the threshold range by 10°C for two consecutive hours, it can be considered as an "overtemperature" anomaly type, and the degree of anomaly is "serious"; by accurately identifying the specific type of environmental problems and their severity, it helps to quickly locate the problem and lay the foundation for formulating reasonable response measures in the next step.
[0076] Step S602, based on a preset rule relationship, determining a corresponding environmental parameter adjustment type and an environmental parameter adjustment index according to the abnormality type and abnormality degree; Specifically, according to the preset rule relationship table, the corresponding environmental parameter adjustment type and adjustment index are pre-configured for each abnormality type and degree of abnormality. The adjustment type refers to which specific environmental parameters need to be changed (such as temperature, humidity, light intensity, etc.), and the adjustment index indicates the magnitude of the adjustment (such as how much percentage to increase / decrease).
[0077] It should be noted that the preset rule relationship table is established based on the principle of combining agricultural science theory with practical experience. These rules are usually derived from the study of specific crop growth requirements, including the range of parameters such as temperature, humidity, and light under optimal growth conditions, and the mechanism of these parameters affecting crop growth. In addition, the establishment of the rule relationship table also needs to consider the urgency of different types of abnormalities and the possible consequences, so as to determine a reasonable response strategy. For example, when the temperature is too high, it may be necessary to take cooling measures immediately to prevent crops from heat damage; when the humidity is slightly low, a milder humidification method can be selected according to the specific situation. Therefore, the formulation of the rule relationship table comprehensively considers the physiological characteristics of crops, the sensitive stages in the growth cycle, and the changing laws of environmental factors, aiming to ensure that the most appropriate actions can be taken when abnormalities occur, maintain an ideal growth environment, and promote the healthy growth of crops.
[0078] Step S603: According to the environmental parameter adjustment type and the environmental parameter adjustment index, the target environmental parameters of the monitoring area are adjusted and a corresponding control strategy is generated.
[0079] Among them, once the type of environmental parameters and adjustment index that need to be adjusted are determined, the system will automatically or semi-automatically issue instructions to related equipment, and the equipment (such as ventilation systems, humidifiers, sunshade nets, etc.) will operate according to the specified parameter adjustment index to achieve the target environmental parameters.
[0080] In the above implementation, environmental parameters are automatically adjusted based on preset rule relationships, thereby achieving effective control of the agricultural environment. This method can not only respond to various environmental anomalies in a timely manner, but also take appropriate countermeasures according to the specific circumstances of the anomalies, thereby effectively improving the crop growth environment and improving the growth quality and yield of crops. At the same time, since the need for manual intervention is reduced, management costs are reduced, and the level of intelligence and sustainability of agricultural production are improved.
[0081] The embodiment of the present application also discloses an agricultural environment information collection and control system.
[0082] An agricultural environmental information collection and control system, the system comprising: Data acquisition module, used to collect remote sensing image information and environmental data information of multiple monitoring areas in the farmland; A preprocessing module is used to preprocess remote sensing image information and environmental data information; A feature extraction module is used to extract features based on remote sensing image information to obtain crop features in each monitoring area; An identification module, used to identify the crop type and crop growth status corresponding to the crop characteristics based on a machine learning algorithm; An abnormal threshold range determination module is used to determine the environmental abnormal threshold range of each monitoring area according to the crop type and crop growth status based on a preset mapping table; The abnormality analysis module is used to perform abnormality analysis on the environmental data information of each monitoring area according to the environmental abnormality threshold range of each monitoring area, and obtain the environmental analysis results of each monitoring area; A control strategy generation module, for generating a control strategy according to the abnormal environment result in response to the environmental analysis result of the monitoring area being an abnormal environment result; The control strategy sending module is used to send the control strategy to the intelligent control system corresponding to the monitoring area.
[0083] An agricultural environment information collection and control system of an embodiment of the present application can implement any of the above-mentioned agricultural environment information collection and control methods, and the specific working process of each module in the agricultural environment information collection and control system can refer to the corresponding process in the above-mentioned method embodiment.
[0084] In the several embodiments provided in this application, it should be understood that the provided methods and systems can be implemented in other ways. For example, the system embodiments described above are only illustrative; for example, the division of a certain module is only a logical function division, and there may be other division methods in actual implementation, such as multiple modules can be combined or integrated into another system, or some features can be ignored or not executed.
[0085] The embodiment of the present application also discloses a computer device.
[0086] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, an agricultural environment information collection and control method as described above is implemented.
[0087] The embodiment of the present application also discloses a computer-readable storage medium.
[0088] A computer-readable storage medium stores a computer program that can be loaded by a processor and execute any one of the agricultural environment information collection and control methods described above.
[0089] Among them, computer-readable storage media can be any tangible medium that contains or stores a program, which can be used by or in conjunction with an instruction execution system, apparatus or device; the program code contained on the computer-readable medium can be transmitted using any appropriate medium, including but not limited to wireless, wire, optical cable, RF, etc., or any suitable combination of the above.
[0090] It should be noted that in the above embodiments, the description of each embodiment has different emphases, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.
[0091] The above are all preferred embodiments of the present application, and are not intended to limit the protection scope of the present application. Any feature disclosed in this specification (including the abstract and drawings), unless otherwise stated, can be replaced by other equivalent or alternative features with similar purposes. That is, unless otherwise stated, each feature is only an example of a series of equivalent or similar features.
Claims
1. A method for collecting and controlling agricultural environmental information, characterized in that: The method comprises: Collect remote sensing image information and environmental data information of multiple monitoring areas in farmland and perform data preprocessing; Extracting features based on the remote sensing image information to obtain crop features in each monitoring area; Based on a machine learning algorithm, identifying the crop type and crop growth status corresponding to the crop characteristics; Based on a preset mapping table, determining an environmental anomaly threshold range for each monitoring area according to the crop type and crop growth status; According to the environmental anomaly threshold range of each monitoring area, the environmental data information of each monitoring area is analyzed for anomalies, and the environmental analysis results of each monitoring area are obtained; In response to the environmental analysis result of the monitoring area being an abnormal environmental result, generating a control strategy according to the abnormal environmental result; The control strategy is sent to the intelligent control system corresponding to the monitoring area.
2. The agricultural environment information collection and control method according to claim 1 is characterized in that: The step of extracting features according to the remote sensing image information to obtain crop features in each monitoring area comprises: Segmenting the image into a plurality of super-pixel regions according to the remote sensing image information, wherein pixels in each super-pixel region have similar features; Extract multispectral features from each superpixel region; Calculate the texture features in each superpixel region based on the gray level co-occurrence matrix; Extract morphological features of each superpixel region; Combining the multispectral features, texture features and morphological features into a feature vector; A key feature subset is selected from the feature vector as crop features based on a feature selection method.
3. The agricultural environment information collection and control method according to claim 2 is characterized in that: Based on the machine learning algorithm, the step of identifying the crop type and crop growth status corresponding to the crop characteristics includes: The crop features are input into a pre-trained crop recognition model to obtain the crop type and crop growth status corresponding to the crop features; wherein the crop recognition model is obtained through iterative training based on multiple sample images and pre-labeled crop type labels and crop growth status labels.
4. The agricultural environment information collection and control method according to claim 1 is characterized in that: The steps of performing abnormal analysis on the environmental data information of each monitoring area according to the environmental abnormality threshold range of each monitoring area include: Determine whether the environmental data information of each monitoring area exceeds the corresponding environmental abnormality threshold range, and if so, determine that the environmental data information of the monitoring area is abnormal; Determine whether the duration of abnormality in the environmental data information of the monitoring area exceeds a preset duration threshold; If it exceeds, the environmental analysis result of the monitoring area is output as an environmental abnormality result, and the corresponding abnormality type and abnormality degree are determined according to the abnormal environmental data information; If it does not exceed, the environmental analysis result of the monitoring area is output as a normal environmental result.
5. The agricultural environment information collection and control method according to claim 4 is characterized in that: The step of generating a control strategy according to the abnormal environmental result comprises: Determine the corresponding abnormality type and abnormality degree according to the environmental abnormality result; Based on the preset rule relationship, determine the corresponding environmental parameter adjustment type and environmental parameter adjustment index according to the abnormality type and abnormality degree; According to the environmental parameter adjustment type and the environmental parameter adjustment index, the target environmental parameters of the monitoring area are adjusted and a corresponding control strategy is generated.
6. The agricultural environment information collection and control method according to claim 3 is characterized in that: The method further comprises a step of training the crop recognition model, wherein the training step comprises: Acquire a sample image data set and perform image preprocessing; wherein the sample image data set includes a plurality of sample images and pre-labeled crop type labels and crop growth status labels; Perform feature extraction on the sample image data set to obtain sample crop features; Divide sample crop features into training set, validation set and test set; Inputting the training set into a pre-built convolutional neural network for initial training; wherein the convolutional neural network includes a convolutional layer, a pooling layer, a fully connected layer and an output layer, and the output layer is used to output the predicted probability of each crop type and crop growth state; The initially trained convolutional neural network is verified and model parameters are optimized based on the verification set and the test set to obtain the trained crop recognition model.
7. The agricultural environment information collection and control method according to claim 6 is characterized in that: The training set is input into a pre-built convolutional neural network for initial training, and the steps of verifying and optimizing model parameters of the initially trained convolutional neural network based on the verification set and the test set include: Inputting the training set into a pre-built convolutional neural network for iterative training, optimizing the model parameters and calculating the cross entropy loss function of the model until the cross entropy loss function meets the preset conditions, thereby obtaining a trained convolutional neural network; Validating the convolutional neural network based on the validation set, evaluating the performance of the convolutional neural network and adjusting the hyperparameters of the model according to the validation results; The prediction ability of the adjusted convolutional neural network is tested based on the test set.
8. An agricultural environmental information collection and control system, characterized in that: The system comprises: Data acquisition module, used to collect remote sensing image information and environmental data information of multiple monitoring areas in the farmland; A preprocessing module is used to preprocess remote sensing image information and environmental data information; A feature extraction module, used to extract features based on the remote sensing image information to obtain crop features in each monitoring area; An identification module, used for identifying the crop type and crop growth status corresponding to the crop characteristics based on a machine learning algorithm; An abnormal threshold range determination module is used to determine the environmental abnormality threshold range of each monitoring area according to the crop type and crop growth status based on a preset mapping table; The abnormality analysis module is used to perform abnormality analysis on the environmental data information of each monitoring area according to the environmental abnormality threshold range of each monitoring area, and obtain the environmental analysis results of each monitoring area; A control strategy generating module, configured to generate a control strategy according to the abnormal environment result in response to the environmental analysis result of the monitoring area being an abnormal environment result; The control strategy sending module is used to send the control strategy to the intelligent control system corresponding to the monitoring area.
9. A computer device, characterized in that: The method comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method according to any one of claims 1 to 7 when executing the program.
10. A computer-readable storage medium, characterized in that: A computer program is stored which can be loaded by a processor and execute the method according to any one of claims 1 to 7.
Citation Information
Cited By
Accurate data acquisition optimization method and system for farmland low-carbon production
CN119338132A
Irrigation area measurement and control integrated gate intelligent scheduling system and method thereof
CN119472353A
Agricultural environment monitoring control method
CN119472893A
Method and system for monitoring growth environment of pearl shellfish
CN120123897A
Visual intelligent agricultural planting control system
CN120255625A
