Adaptive crop environment monitoring system and method based on deep learning

Abstract

The invention relates to the technical field of environment monitoring and early warning, and discloses a self-adaptive crop environment monitoring system and method based on deep learning, and the system comprises an acquisition module which is configured to judge whether to carry out abnormal environment early warning on a to-be-monitored area based on real-time crop environment data and crop image data; the determination module is configured to construct a historical crop image data set and determine an initial early warning value of the to-be-monitored area according to the historical crop image data set; the judgment module is configured to judge whether to adjust the initial early warning value or not based on the analysis result; the processing module is configured to adjust the initial early warning value according to the historical crop environment data set to obtain a final early warning value; and the early warning module is configured to determine the early warning level of the to-be-monitored area according to the final early warning value. According to the invention, adaptive monitoring of the crop environment can be realized, and the accuracy and timeliness of crop environment monitoring are improved.

Description

Technical Field

[0001] This invention relates to the field of environmental monitoring and early warning technology, and more specifically, to an adaptive crop environment monitoring system and method based on deep learning. Background Technology

[0002] In greenhouses where crops are grown, environmental monitoring is crucial, directly impacting crop growth and final yield. Traditional environmental monitoring methods often rely on fixed threshold settings, lacking adaptability to environmental changes. When environmental conditions fluctuate significantly, fixed thresholds cannot accurately reflect the actual situation, easily leading to misjudgments or missed detections. For example, under certain climatic conditions, parameters such as temperature and humidity inside the greenhouse may experience brief abnormal fluctuations, but these fluctuations may not have a substantial impact on crop growth. However, using traditional fixed threshold monitoring methods may trigger unnecessary warnings, causing unnecessary trouble for growers.

[0003] Therefore, it is necessary to design an adaptive crop environment monitoring system and method based on deep learning to solve the problems existing in the current technology. Summary of the Invention

[0004] In view of this, the present invention proposes an adaptive crop environment monitoring system and method based on deep learning, which aims to solve the problem that fixed thresholds in the current technology cannot reflect the actual situation in a timely and accurate manner, and are prone to misjudgment or omission.

[0005] In one aspect, this invention proposes a deep learning-based adaptive crop environment monitoring system, comprising: The acquisition module is configured to determine the area to be monitored, acquire real-time crop environment data of the area to be monitored and crop image data of the crop to be monitored located in the area to be monitored, and determine whether to issue an abnormal environment warning for the area to be monitored based on the real-time crop environment data and crop image data. The determination module is configured to, when it is determined that an abnormal environment warning should be issued for the area to be monitored, collect historical crop image data corresponding to the previous n collection intervals of the crop to be monitored, construct a historical crop image dataset, and determine the initial warning value of the area to be monitored based on the historical crop image dataset; The judgment module is configured to collect the status information of the sensors located inside the area to be monitored, parse the status information, and determine whether to adjust the initial warning value based on the parsing result; The processing module is configured to, when it is determined that the initial warning value needs to be adjusted, collect historical environmental data corresponding to the previous n collection intervals of the sensor, construct a historical crop environment dataset, adjust the initial warning value according to the historical crop environment dataset, and obtain the final warning value. The early warning module is configured to determine the early warning level of the area to be monitored based on the final early warning value.

[0006] Furthermore, when determining whether to issue an abnormal environmental warning for the monitored area based on the real-time crop environment data and crop image data, the following steps are included: The real-time crop environment data is parsed to obtain real-time crop environment feature values; The crop image data is parsed to obtain crop image anomalies; Obtain the correlation coefficient between the real-time crop environment feature values ​​and the crop image anomalies. The correlation coefficient is used to determine whether to issue an abnormal environmental warning for the area to be monitored.

[0007] Furthermore, when obtaining the correlation coefficient between the real-time crop environment feature values ​​and the crop image anomalies, the following steps are included: The real-time crop environment feature values ​​and crop image outliers are normalized to obtain normalized environment feature values ​​and normalized crop image outliers. A correlation coefficient calculation model is constructed using a deep learning algorithm, and the normalized environmental feature value and the normalized crop image outlier value are input into the correlation coefficient calculation model. Obtain the output value of the correlation coefficient calculation model, and denote it as the correlation coefficient.

[0008] Furthermore, when determining whether to issue an abnormal environmental warning for the monitored area based on the correlation coefficient, the following steps are included: The correlation coefficient is compared with the correlation coefficient threshold, and the comparison result is used to determine whether to issue an abnormal environment warning for the area to be monitored. When the correlation coefficient is greater than or equal to the correlation coefficient threshold, it is determined that an abnormal environment warning should be issued for the area to be monitored. Otherwise, no abnormal environmental warning will be issued for the area to be monitored.

[0009] Further, when determining the initial warning value for the area to be monitored based on the historical crop image dataset, the process includes: The abnormal change rate and abnormal change range of the crop to be monitored are determined based on the historical crop image dataset. Construct an abnormal change vector group based on the abnormal change rate and abnormal change range values; The abnormal change vector group is compared with the historical abnormal change group, and the initial warning value of the area to be monitored is determined based on the comparison result. If there is a historical abnormal change vector group in the historical abnormal change group that is the same as the abnormal change vector group, the historical warning value corresponding to the historical abnormal change vector group shall be used as the initial warning value. If there is no historical abnormal change vector group in the historical abnormal change group that is the same as the abnormal change vector group, the initial warning value is determined based on the abnormal change vector group.

[0010] Further, when determining the initial warning value based on the abnormal change vector group, the process includes: The abnormal change rate is compared with the abnormal change rate threshold, and the abnormal change range value is compared with the abnormal change range threshold. The initial warning value is determined based on the comparison results. When the abnormal change rate is greater than or equal to the abnormal change rate threshold, and the abnormal change range value is greater than or equal to the abnormal change range threshold, the initial warning value is determined to be the first warning value. When the abnormal change rate is greater than or equal to the abnormal change rate threshold, and the abnormal change range value is less than the abnormal change range threshold, the initial warning value is determined to be the second warning value. When the abnormal change rate is less than the abnormal change rate threshold and the abnormal change range value is greater than or equal to the abnormal change range threshold, the initial warning value is determined to be the third warning value. When the abnormal change rate is less than the abnormal change rate threshold and the abnormal change range value is less than the abnormal change range threshold, the initial warning value is determined to be the fourth warning value.

[0011] Furthermore, when parsing the status information and determining whether to adjust the initial warning value based on the parsing result, the process includes: The status information is analyzed to determine whether any sensors are malfunctioning. If so, identify the abnormal sensor, analyze the type and degree of abnormality of the sensor, and determine the potential impact value of the sensor on the initial warning value by combining the weight of the real-time crop environment data collected by the sensor in all the real-time crop environment data. Determine whether to adjust the initial warning value based on the potential impact value; When the potential impact value is less than or equal to a%, it is determined that the initial warning value will not be adjusted. Otherwise, it is determined that the initial warning value should be adjusted.

[0012] Furthermore, when adjusting the initial warning value based on the historical crop environment dataset to obtain the final warning value, the following steps are included: Remove real-time crop environment data collected by sensors with abnormalities from the historical crop environment dataset to construct a normal crop environment dataset; The normal crop environment dataset is parsed to obtain the crop environment fluctuation characteristic values ​​of the area to be monitored; The crop environment fluctuation characteristic value is compared with the first crop environment fluctuation characteristic value and the second crop environment fluctuation characteristic value, and the adjustment coefficient of the initial warning value is determined according to the comparison result; wherein, the first crop environment fluctuation characteristic value is smaller than the second crop environment fluctuation characteristic value; When the crop environment fluctuation characteristic value is less than or equal to the first crop environment fluctuation characteristic value, the initial warning value is determined to be the first adjustment coefficient; When the crop environment fluctuation characteristic value is greater than the first crop environment fluctuation characteristic value and less than or equal to the second crop environment fluctuation characteristic value, the initial warning value is determined to be the second adjustment coefficient. When the crop environment fluctuation characteristic value is greater than the second crop environment fluctuation characteristic value, the initial warning value is determined to be the third adjustment coefficient; The product of the adjustment coefficient and the initial warning value is used as the final warning value.

[0013] Furthermore, when determining the warning level of the area to be monitored based on the final warning value, the following steps are included: The final warning value is compared with a preset warning level mapping table, and the warning level is determined based on the comparison result.

[0014] Compared with existing technologies, the beneficial effects of this invention are as follows: The deep learning-based adaptive crop environment monitoring system provided by this invention can achieve adaptive monitoring of the crop environment, improving the accuracy and timeliness of crop environment monitoring. Traditional environmental monitoring systems often lack the ability to adapt to environmental changes and struggle to dynamically adjust warning values ​​and warning levels according to actual conditions. This system, however, uses deep learning algorithms combined with real-time crop environment data and crop image data to accurately determine whether the environment is abnormal and dynamically adjusts warning values ​​based on historical data and sensor status information, thereby improving the accuracy and reliability of warnings. Furthermore, the system has strong flexibility and scalability. The acquisition module can determine the monitoring area according to different monitoring needs and collect corresponding real-time crop environment data and crop image data. The various modules collaborate to form a complete monitoring system that can adapt to different environmental monitoring scenarios. Moreover, the deep learning algorithm used in the system can be optimized and improved according to actual conditions to better adapt to constantly changing environmental monitoring needs.

[0015] In another aspect, this invention also proposes a deep learning-based adaptive crop environment monitoring method, comprising the following steps: The monitoring area is determined, real-time crop environment data of the monitoring area and crop image data of the crops to be monitored located in the monitoring area are collected, and based on the real-time crop environment data and crop image data, it is determined whether to issue an abnormal environment warning for the monitoring area. When it is determined that an abnormal environment warning should be issued for the area to be monitored, historical crop image data corresponding to the previous n collection intervals of the crop to be monitored are collected, and a historical crop image dataset is constructed. The initial warning value of the area to be monitored is determined based on the historical crop image dataset. Collect status information from sensors located within the area to be monitored, analyze the status information, and determine whether to adjust the initial warning value based on the analysis results; When it is determined that the initial warning value needs to be adjusted, historical environmental data corresponding to the previous n acquisition intervals of the sensor are collected, and a historical crop environment dataset is constructed. The initial warning value is adjusted according to the historical crop environment dataset to obtain the final warning value. The warning level of the area to be monitored is determined based on the final warning value.

[0016] It is understandable that the aforementioned deep learning-based adaptive crop environment monitoring system and method have the same beneficial effects, and will not be elaborated further here. Attached Figure Description

[0017] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 A structural block diagram of a deep learning-based adaptive crop environment monitoring system provided in an embodiment of the present invention; Figure 2 A flowchart of a deep learning-based adaptive crop environment monitoring method provided in an embodiment of the present invention. Detailed Implementation

[0018] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the disclosure to those skilled in the art. It should be noted that, unless otherwise specified, embodiments and features in the embodiments of the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0019] See Figure 1 As shown in some embodiments of this application, this embodiment provides a deep learning-based adaptive crop environment monitoring system, including: The acquisition module is configured to determine the area to be monitored, acquire real-time crop environment data of the area to be monitored and crop image data of the crop to be monitored located in the area to be monitored, and determine whether to issue an abnormal environment warning for the area to be monitored based on the real-time crop environment data and crop image data. The determination module is configured to, when it is determined that an abnormal environment warning should be issued for the area to be monitored, collect historical crop image data corresponding to the previous n collection intervals of the crop to be monitored, construct a historical crop image dataset, and determine the initial warning value of the area to be monitored based on the historical crop image dataset; The judgment module is configured to collect the status information of the sensors located inside the area to be monitored, parse the status information, and determine whether to adjust the initial warning value based on the parsing result; The processing module is configured to, when it is determined that the initial warning value needs to be adjusted, collect historical environmental data corresponding to the previous n collection intervals of the sensor, construct a historical crop environment dataset, adjust the initial warning value according to the historical crop environment dataset, and obtain the final warning value. The early warning module is configured to determine the early warning level of the area to be monitored based on the final early warning value.

[0020] In this embodiment, n is preferably 3, meaning that historical crop image data corresponding to the first 3 acquisition intervals of the crop to be monitored and historical environmental data corresponding to the first 3 acquisition intervals of the sensor are collected. Choosing n as 3 comprehensively considers both the timeliness and accuracy of the data. The acquisition interval is preferably 1 hour. Such an acquisition interval ensures that enough data is obtained for analysis, without data redundancy due to too short an interval, and without data loss of timeliness due to too long an interval.

[0021] In this embodiment, the area to be monitored is preferably farmland or crop greenhouse.

[0022] It is understood that the deep learning-based adaptive crop environment monitoring system provided in this embodiment can achieve adaptive monitoring of the crop environment, improving the accuracy and timeliness of crop environment monitoring. Traditional environmental monitoring systems often lack the ability to adapt to environmental changes and struggle to dynamically adjust warning values ​​and levels according to actual conditions. This system, however, uses deep learning algorithms combined with real-time crop environment data and crop image data to accurately determine whether the environment is abnormal and dynamically adjusts warning values ​​based on historical data and sensor status information, thereby improving the accuracy and reliability of warnings. Furthermore, the system has strong flexibility and scalability. The acquisition module can determine the monitoring area according to different monitoring needs and collect corresponding real-time crop environment data and crop image data. The various modules collaborate to form a complete monitoring system that can adapt to different environmental monitoring scenarios. Moreover, the deep learning algorithms used in the system can be optimized and improved according to actual conditions to better adapt to constantly changing environmental monitoring needs.

[0023] Specifically, when determining whether to issue an abnormal environmental warning for the monitored area based on the real-time crop environment data and crop image data, the following steps are included: The real-time crop environment data is parsed to obtain real-time crop environment feature values; The crop image data is parsed to obtain crop image anomalies; Obtain the correlation coefficient between the real-time crop environment feature values ​​and the crop image anomalies. The correlation coefficient is used to determine whether to issue an abnormal environmental warning for the area to be monitored.

[0024] Understandably, real-time crop environmental characteristics include real-time measurements of environmental parameters such as temperature, humidity, light intensity, wind speed, and air pressure. These parameters directly reflect the current environmental conditions of the monitored area. Crop image anomalies reflect the differences between the monitored crop's growth status and its normal state, manifested as changes in leaf color, abnormal leaf shape, and the degree of pest and disease infection. By analyzing the correlation coefficient between real-time crop environmental characteristics and crop image anomalies, we can comprehensively consider environmental factors and crop growth status, and more accurately determine whether an abnormal environmental warning is needed.

[0025] Specifically, obtaining the correlation coefficient between the real-time crop environment feature values ​​and the crop image anomalies includes: The real-time crop environment feature values ​​and crop image outliers are normalized to obtain normalized environment feature values ​​and normalized crop image outliers. A correlation coefficient calculation model is constructed using a deep learning algorithm, and the normalized environmental feature value and the normalized crop image outlier value are input into the correlation coefficient calculation model. Obtain the output value of the correlation coefficient calculation model, and denote it as the correlation coefficient.

[0026] The construction process of the correlation coefficient calculation model can be understood as follows: First, a large amount of historical data is collected, including real-time crop environmental feature values ​​under different environmental conditions and corresponding crop image anomalies. Next, the collected historical data is divided into training, validation, and test sets. The training set is used to train the model, the validation set is used to adjust the model's hyperparameters, and the test set is used to evaluate the model's performance. Then, a suitable deep learning algorithm architecture, such as a Convolutional Neural Network (CNN) or a Recurrent Neural Network (RNN), is selected as the basic architecture for the correlation coefficient calculation model. CNNs are suitable for processing image data and can extract features from images; RNNs are better at processing sequence data and can be used to capture the time-series features of environmental data. After determining the basic architecture, the model is initialized by setting its initial parameters. Subsequently, the model is trained using the training set, and the model's parameters are continuously adjusted to make the model's output as close as possible to the actual correlation coefficient. During training, a loss function is used to measure the difference between the model's output and the actual value, and optimization algorithms (such as stochastic gradient descent) are used to update the model's parameters to minimize the loss function. During training, the model needs to be validated using a validation set. Based on the validation results, hyperparameters such as learning rate and batch size are adjusted to improve the model's generalization ability. Once the model's performance on the validation set reaches a satisfactory level, a test set is used for final evaluation to ensure good performance on unseen data. Through this training and adjustment process, a model capable of accurately calculating the correlation coefficient between real-time crop environmental features and crop image anomalies is obtained. This model can output the corresponding correlation coefficient based on the input normalized environmental features and normalized crop image anomalies, providing an important basis for determining whether to issue an anomaly warning for the monitored area.

[0027] Specifically, when determining whether to issue an abnormal environmental warning for the monitored area based on the correlation coefficient, the following steps are included: The correlation coefficient is compared with the correlation coefficient threshold, and the comparison result is used to determine whether to issue an abnormal environment warning for the area to be monitored. When the correlation coefficient is greater than or equal to the correlation coefficient threshold, it is determined that an abnormal environment warning should be issued for the area to be monitored. Otherwise, no abnormal environmental warning will be issued for the area to be monitored.

[0028] Understandably, setting the correlation coefficient threshold is crucial, requiring the integration of extensive historical data and practical environmental monitoring experience. In practical applications, the correlation coefficient threshold can be flexibly adjusted based on different monitored areas and needs to balance the accuracy and timeliness of early warnings. For example, when monitoring crop-growing areas that are more sensitive to environmental changes, the correlation coefficient threshold can be appropriately lowered to detect environmental anomalies more promptly; conversely, when monitoring crop-growing areas with stronger environmental adaptability, the correlation coefficient threshold can be appropriately increased to reduce unnecessary warnings.

[0029] Specifically, determining the initial warning value for the area to be monitored based on the historical crop image dataset includes: The abnormal change rate and abnormal change range of the crop to be monitored are determined based on the historical crop image dataset. Construct an abnormal change vector group based on the abnormal change rate and abnormal change range values; The abnormal change vector group is compared with the historical abnormal change group, and the initial warning value of the area to be monitored is determined based on the comparison result. If there is a historical abnormal change vector group in the historical abnormal change group that is the same as the abnormal change vector group, the historical warning value corresponding to the historical abnormal change vector group shall be used as the initial warning value. If there is no historical abnormal change vector group in the historical abnormal change group that is the same as the abnormal change vector group, the initial warning value is determined based on the abnormal change vector group.

[0030] Understandably, the rate of abnormal change reflects the speed at which the monitored crop's growth status changes abnormally over a period of time, while the range of abnormal change reflects the magnitude of this change. By constructing an abnormal change vector set from the rate and range of abnormal change, the abnormal change characteristics of the monitored crop can be described more comprehensively and accurately. The historical abnormal change set is accumulated by the system over a long period of monitoring and contains historical warning values ​​corresponding to various abnormal change characteristics. When a historical abnormal change vector set exists that is identical to the current one, its corresponding historical warning value is directly used as the initial warning value, thus fully utilizing historical experience and improving the accuracy of the warning. However, when no identical historical abnormal change vector set exists, it is necessary to predict and determine the initial warning value based on the current abnormal change vector set, combined with a deep learning algorithm.

[0031] Specifically, determining the initial warning value based on the abnormal change vector group includes: The abnormal change rate is compared with the abnormal change rate threshold, and the abnormal change range value is compared with the abnormal change range threshold. The initial warning value is determined based on the comparison results. When the abnormal change rate is greater than or equal to the abnormal change rate threshold, and the abnormal change range value is greater than or equal to the abnormal change range threshold, the initial warning value is determined to be the first warning value. When the abnormal change rate is greater than or equal to the abnormal change rate threshold, and the abnormal change range value is less than the abnormal change range threshold, the initial warning value is determined to be the second warning value. When the abnormal change rate is less than the abnormal change rate threshold and the abnormal change range value is greater than or equal to the abnormal change range threshold, the initial warning value is determined to be the third warning value. When the abnormal change rate is less than the abnormal change rate threshold and the abnormal change range value is less than the abnormal change range threshold, the initial warning value is determined to be the fourth warning value.

[0032] Understandably, the initial warning values ​​are ranked in the following order: First Warning Value > Second Warning Value > Third Warning Value > Fourth Warning Value. This setting of warning values ​​is based on the severity of abnormal changes in the monitored crop. When both the rate of abnormal change and the range of abnormal change exceed their respective thresholds, it indicates the most severe abnormality in crop growth. In this case, the highest first warning value is set to promptly alert relevant personnel to take appropriate countermeasures and avoid potential serious losses. When the rate of abnormal change exceeds the threshold but the range of abnormal change does not, or vice versa, the second and third warning values ​​are respectively assigned, indicating a certain degree of abnormality in crop growth, but less severe than the first scenario. When neither the rate of abnormal change nor the range of abnormal change exceeds the threshold, the fourth warning value is assigned. At this point, the degree of abnormality in crop growth is relatively low, but continued monitoring is still necessary.

[0033] Specifically, when parsing the status information and determining whether to adjust the initial warning value based on the parsing result, the process includes: The status information is analyzed to determine whether any sensors are malfunctioning. If so, identify the abnormal sensor, analyze the type and degree of abnormality of the sensor, and determine the potential impact value of the sensor on the initial warning value by combining the weight of the real-time crop environment data collected by the sensor in all the real-time crop environment data. Determine whether to adjust the initial warning value based on the potential impact value; When the potential impact value is less than or equal to a%, it is determined that the initial warning value will not be adjusted. Otherwise, it is determined that the initial warning value should be adjusted.

[0034] It's understandable that sensor malfunctions stem from hardware failures, such as damaged sensing elements or loose circuit connections, leading to inaccurate or completely erroneous data. Alternatively, the sensor may be affected by external environmental factors. For instance, strong electromagnetic interference can cause data transmission deviations, while extreme conditions like high temperature and humidity can impact sensor performance, resulting in abnormal real-time crop environment data. Once a sensor malfunction is identified, analyzing the type of anomaly is crucial. Hardware failures are often severe, potentially rendering the sensor completely inoperable and the data unusable. External interference, on the other hand, requires assessment of its severity based on the specific interference. Weak interference may cause minor fluctuations, while strong interference can lead to significant data deviations.

[0035] In this embodiment, the calculation process for the potential impact value is as follows: First, the weight of the real-time crop environment data collected by the sensor exhibiting anomalies among all real-time crop environment data is determined. This weight reflects its importance to the overall environmental monitoring results and can be determined comprehensively through historical data analysis, the sensor's position and role in the monitoring system, and other factors. Next, the type and degree of anomaly of the sensor are evaluated. For example, a hardware failure with completely damaged sensing elements is assigned an anomaly degree of 1, while a slight loosening of the circuit connection is assigned 0.7. External environmental interference with data deviation exceeding the range under strong electromagnetic interference is assigned 0.9, and minor data fluctuations under normal environmental fluctuations are assigned 0.3. Then, coefficients are assigned according to the anomaly type and degree, with higher anomalies assigned larger coefficients and lower anomalies assigned smaller coefficients. For example, a high anomaly coefficient for hardware failure is assigned 0.8, and a low anomaly coefficient for external environmental interference is assigned 0.2. Finally, the weight of the sensor's real-time crop environment data is multiplied by the anomaly degree, and then multiplied by the anomaly degree coefficient to obtain the potential impact value.

[0036] Specifically, when adjusting the initial warning value based on the historical crop environment dataset to obtain the final warning value, the following steps are included: Remove real-time crop environment data collected by sensors with abnormalities from the historical crop environment dataset to construct a normal crop environment dataset; The normal crop environment dataset is parsed to obtain the crop environment fluctuation characteristic values ​​of the area to be monitored; The crop environment fluctuation characteristic value is compared with the first crop environment fluctuation characteristic value and the second crop environment fluctuation characteristic value, and the adjustment coefficient of the initial warning value is determined according to the comparison result; wherein, the first crop environment fluctuation characteristic value is smaller than the second crop environment fluctuation characteristic value; When the crop environment fluctuation characteristic value is less than or equal to the first crop environment fluctuation characteristic value, the initial warning value is determined to be the first adjustment coefficient; When the crop environment fluctuation characteristic value is greater than the first crop environment fluctuation characteristic value and less than or equal to the second crop environment fluctuation characteristic value, the initial warning value is determined to be the second adjustment coefficient. When the crop environment fluctuation characteristic value is greater than the second crop environment fluctuation characteristic value, the initial warning value is determined to be the third adjustment coefficient; The product of the adjustment coefficient and the initial warning value is used as the final warning value.

[0037] Understandably, crop environmental fluctuation characteristic values ​​are quantitative indicators reflecting the stability and changing trends of the environment in the monitored area. These include, for example, the standard deviation of environmental parameters (temperature, humidity, light intensity, etc.) over a period of time (a larger deviation indicates more severe environmental fluctuations), or the rate of change of environmental parameters (a larger rate indicates more rapid environmental changes). The first and second crop environmental fluctuation characteristic values ​​are key thresholds determined based on historical environmental data and monitoring experience. The first crop environmental fluctuation characteristic value represents the boundary of relative environmental stability. When the crop environmental fluctuation characteristic value is less than or equal to this value, the environment is relatively stable. The first adjustment coefficient is used to adjust the initial warning value, reducing warning sensitivity and avoiding unnecessary warnings. The second crop environmental fluctuation characteristic value represents the boundary of severe environmental fluctuations. When the crop environmental fluctuation characteristic value is greater than this value, environmental changes are significant and may affect crop growth. The third adjustment coefficient is used to adjust the initial warning value, increasing warning sensitivity and enabling timely detection of environmental anomalies. When the crop environmental fluctuation characteristic value falls between these two values, the environment experiences some fluctuations but not severe ones. The second adjustment coefficient is used to adjust the warning sensitivity to a moderate level. Through this adjustment method, warning values ​​can be flexibly adjusted according to actual environmental fluctuations, improving the accuracy and effectiveness of warnings and providing reliable guarantees for environmental management and crop growth.

[0038] Specifically, determining the warning level of the area to be monitored based on the final warning value includes: The final warning value is compared with a preset warning level mapping table, and the warning level is determined based on the comparison result.

[0039] Understandably, the pre-defined warning level mapping table is a pre-set correspondence table that clearly defines the warning level corresponding to different final warning value ranges. Its formulation is based on extensive historical data and professional agricultural environmental monitoring experience. Warning levels are typically divided into low, medium, and high basic levels, and can be further subdivided to accurately reflect the severity of environmental anomalies in the monitored area. When the final warning value is in a low range, corresponding to a "low warning level," it indicates that while there is an environmental anomaly, it is relatively minor and has little potential impact on crop growth; relevant personnel can strengthen daily monitoring. In a medium range, corresponding to a "medium warning level," it means that the degree of environmental anomaly has increased and has already affected crop growth, requiring proactive measures to mitigate it. In a high range, corresponding to a "high warning level," it indicates that the environmental anomaly is severe and may threaten crop growth or even lead to reduced yields or crop failure; immediate emergency measures must be taken to reduce losses. By comparing the final warning value with the mapping table, the warning level can be quickly and accurately determined, providing a basis for decision-making for relevant personnel and ensuring the stability of crop growth and agricultural production.

[0040] See Figure 2 As shown in some embodiments of this application, this embodiment provides a deep learning-based adaptive crop environment monitoring method, including the following steps: S100: Determine the area to be monitored, collect real-time crop environment data of the area to be monitored and crop image data of the crops to be monitored located in the area to be monitored, and determine whether to issue an abnormal environment warning for the area to be monitored based on the real-time crop environment data and crop image data. S200: When it is determined that an abnormal environment warning should be issued for the area to be monitored, historical crop image data corresponding to the previous n collection intervals of the crop to be monitored are collected, and a historical crop image dataset is constructed. The initial warning value of the area to be monitored is determined based on the historical crop image dataset. S300: Collect the status information of the sensors located inside the area to be monitored, analyze the status information, and determine whether to adjust the initial warning value based on the analysis result; S400: When it is determined that the initial warning value needs to be adjusted, historical environmental data corresponding to the previous n acquisition intervals of the sensor are collected, and a historical crop environment dataset is constructed. The initial warning value is adjusted according to the historical crop environment dataset to obtain the final warning value. S500: Determine the warning level of the area to be monitored based on the final warning value.

[0041] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program goods. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program goods embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0042] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program goods according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0043] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0044] These computer program instructions can also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0045] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A deep learning-based adaptive crop environment monitoring system, characterized in that, include: The acquisition module is configured to determine the area to be monitored, acquire real-time crop environment data of the area to be monitored and crop image data of the crop to be monitored located in the area to be monitored, and determine whether to issue an abnormal environment warning for the area to be monitored based on the real-time crop environment data and crop image data. The determination module is configured to, when it is determined that an abnormal environment warning should be issued for the area to be monitored, collect historical crop image data corresponding to the previous n collection intervals of the crop to be monitored, construct a historical crop image dataset, and determine the initial warning value of the area to be monitored based on the historical crop image dataset; The judgment module is configured to collect the status information of the sensors located inside the area to be monitored, parse the status information, and determine whether to adjust the initial warning value based on the parsing result; The processing module is configured to, when it is determined that the initial warning value needs to be adjusted, collect historical environmental data corresponding to the previous n collection intervals of the sensor, construct a historical crop environment dataset, adjust the initial warning value according to the historical crop environment dataset, and obtain the final warning value. The early warning module is configured to determine the early warning level of the area to be monitored based on the final early warning value.

2. The deep learning-based adaptive crop environment monitoring system according to claim 1, characterized in that, When determining whether to issue an abnormal environmental warning for the monitored area based on the real-time crop environment data and crop image data, the following are included: The real-time crop environment data is parsed to obtain real-time crop environment feature values; The crop image data is parsed to obtain crop image anomalies; Obtain the correlation coefficient between the real-time crop environment feature values ​​and the crop image anomalies. The correlation coefficient is used to determine whether to issue an abnormal environmental warning for the area to be monitored.

3. The deep learning-based adaptive crop environment monitoring system according to claim 2, characterized in that, When obtaining the correlation coefficient between the real-time crop environment feature values ​​and the crop image anomalies, the following steps are included: The real-time crop environment feature values ​​and crop image outliers are normalized to obtain normalized environment feature values ​​and normalized crop image outliers. A correlation coefficient calculation model is constructed using a deep learning algorithm, and the normalized environmental feature value and the normalized crop image outlier value are input into the correlation coefficient calculation model. Obtain the output value of the correlation coefficient calculation model, and denote it as the correlation coefficient.

4. The deep learning-based adaptive crop environment monitoring system according to claim 3, characterized in that, When determining whether to issue an abnormal environmental warning for the monitored area based on the correlation coefficient, the following are included: The correlation coefficient is compared with the correlation coefficient threshold, and the comparison result is used to determine whether to issue an abnormal environment warning for the area to be monitored. When the correlation coefficient is greater than or equal to the correlation coefficient threshold, it is determined that an abnormal environment warning should be issued for the area to be monitored. Otherwise, no abnormal environmental warning will be issued for the area to be monitored.

5. The deep learning-based adaptive crop environment monitoring system according to claim 4, characterized in that, When determining the initial warning value for the area to be monitored based on the historical crop image dataset, the following are included: The abnormal change rate and abnormal change range of the crop to be monitored are determined based on the historical crop image dataset. Construct an abnormal change vector group based on the abnormal change rate and abnormal change range values; The abnormal change vector group is compared with the historical abnormal change group, and the initial warning value of the area to be monitored is determined based on the comparison result. If there is a historical abnormal change vector group in the historical abnormal change group that is the same as the abnormal change vector group, the historical warning value corresponding to the historical abnormal change vector group shall be used as the initial warning value. If there is no historical abnormal change vector group in the historical abnormal change group that is the same as the abnormal change vector group, the initial warning value is determined based on the abnormal change vector group.

6. The deep learning-based adaptive crop environment monitoring system according to claim 5, characterized in that, When determining the initial warning value based on the abnormal change vector group, the following is included: The abnormal change rate is compared with the abnormal change rate threshold, and the abnormal change range value is compared with the abnormal change range threshold. The initial warning value is determined based on the comparison results. When the abnormal change rate is greater than or equal to the abnormal change rate threshold, and the abnormal change range value is greater than or equal to the abnormal change range threshold, the initial warning value is determined to be the first warning value. When the abnormal change rate is greater than or equal to the abnormal change rate threshold, and the abnormal change range value is less than the abnormal change range threshold, the initial warning value is determined to be the second warning value. When the abnormal change rate is less than the abnormal change rate threshold and the abnormal change range value is greater than or equal to the abnormal change range threshold, the initial warning value is determined to be the third warning value. When the abnormal change rate is less than the abnormal change rate threshold and the abnormal change range value is less than the abnormal change range threshold, the initial warning value is determined to be the fourth warning value.

7. The deep learning-based adaptive crop environment monitoring system according to claim 6, characterized in that, When parsing the status information and determining whether to adjust the initial warning value based on the parsing result, the process includes: The status information is analyzed to determine whether any sensors are malfunctioning. If so, identify the abnormal sensor, analyze the type and degree of abnormality of the sensor, and determine the potential impact value of the sensor on the initial warning value by combining the weight of the real-time crop environment data collected by the sensor in all the real-time crop environment data. Determine whether to adjust the initial warning value based on the potential impact value; When the potential impact value is less than or equal to a%, it is determined that the initial warning value will not be adjusted. Otherwise, it is determined that the initial warning value should be adjusted.

8. The deep learning-based adaptive crop environment monitoring system according to claim 7, characterized in that, When adjusting the initial warning value based on the historical crop environment dataset to obtain the final warning value, the following steps are included: Remove real-time crop environment data collected by sensors with abnormalities from the historical crop environment dataset to construct a normal crop environment dataset; The normal crop environment dataset is parsed to obtain the crop environment fluctuation characteristic values ​​of the area to be monitored; The crop environment fluctuation characteristic value is compared with the first crop environment fluctuation characteristic value and the second crop environment fluctuation characteristic value, and the adjustment coefficient of the initial warning value is determined according to the comparison result; wherein, the first crop environment fluctuation characteristic value is smaller than the second crop environment fluctuation characteristic value; When the crop environment fluctuation characteristic value is less than or equal to the first crop environment fluctuation characteristic value, the initial warning value is determined to be the first adjustment coefficient; When the crop environment fluctuation characteristic value is greater than the first crop environment fluctuation characteristic value and less than or equal to the second crop environment fluctuation characteristic value, the initial warning value is determined to be the second adjustment coefficient. When the crop environment fluctuation characteristic value is greater than the second crop environment fluctuation characteristic value, the initial warning value is determined to be the third adjustment coefficient; The product of the adjustment coefficient and the initial warning value is used as the final warning value.

9. The deep learning-based adaptive crop environment monitoring system according to claim 8, characterized in that, When determining the warning level of the area to be monitored based on the final warning value, the following are included: The final warning value is compared with a preset warning level mapping table, and the warning level is determined based on the comparison result.

10. A deep learning-based adaptive crop environment monitoring method, applied to the deep learning-based adaptive crop environment monitoring system as described in any one of claims 1-9, characterized in that, include: The monitoring area is determined, real-time crop environment data of the monitoring area and crop image data of the crops to be monitored located in the monitoring area are collected, and based on the real-time crop environment data and crop image data, it is determined whether to issue an abnormal environment warning for the monitoring area. When it is determined that an abnormal environment warning should be issued for the area to be monitored, historical crop image data corresponding to the previous n collection intervals of the crop to be monitored are collected, and a historical crop image dataset is constructed. The initial warning value of the area to be monitored is determined based on the historical crop image dataset. Collect status information from sensors located within the area to be monitored, analyze the status information, and determine whether to adjust the initial warning value based on the analysis results; When it is determined that the initial warning value needs to be adjusted, historical environmental data corresponding to the previous n acquisition intervals of the sensor are collected, and a historical crop environment dataset is constructed. The initial warning value is adjusted according to the historical crop environment dataset to obtain the final warning value. The warning level of the area to be monitored is determined based on the final warning value.

Images