Iot-based intelligent agricultural product pest and disease diagnosis system and method

Abstract

This invention discloses an IoT-based smart agricultural product pest and disease diagnosis system and method, belonging to the field of smart agricultural product technology. To address the problem of the wide variety of crop pests and diseases, and the potential for poor identification of rare or newly emerging pests and diseases, the IoT-based smart agricultural product pest and disease diagnosis system includes an image acquisition unit, a feature extraction unit, a diagnostic decision-making unit, and a remote interaction unit. The image acquisition unit uses cameras and sensors deployed in the farmland to capture real-time image data of the crop growth environment and pests and diseases, ensuring timely diagnosis. The feature extraction unit establishes a pest and disease feature database and constructs a key feature recognition model, achieving intelligent processing of pest and disease images, identifying and diagnosing multiple pest and disease types. The remote interaction unit improves user convenience, reduces the cost of manual inspection and diagnosis, increases agricultural production efficiency, and improves crop yield and quality.

Description

Technical Field

[0001] This invention relates to the field of smart agricultural product technology, and in particular to a smart agricultural product pest and disease diagnosis system and method based on the Internet of Things. Background Technology

[0002] Traditional agricultural products face numerous problems in pest and disease control, such as resource waste, environmental pollution, and low production efficiency. Traditional pest and disease monitoring methods often rely on manual inspections and experience-based judgment, which are not only time-consuming and labor-intensive but also prone to misjudgment. With the rapid development of IoT technology, smart agriculture has become an important approach to solving these problems.

[0003] Patent application CN116524321A discloses an intelligent diagnostic method and system for crop diseases and pests, including the following steps: acquiring images and feature information of diseases and pests; preprocessing the images to obtain preprocessed images; inputting the preprocessed images and feature information into a trained crop disease and pest diagnosis model; and obtaining diagnostic results through the model. By preprocessing the acquired images and then using the trained model to diagnose the preprocessed images and feature information, a standardized and precise diagnosis of crop diseases and pests is achieved, effectively preventing such diseases and pests and providing key practical guidance for users to further implement disease and pest control measures.

[0004] While the aforementioned patents utilize technologies such as IoT sensors, big data analytics, and cloud computing to achieve real-time monitoring and data analysis of crop growth environments, providing a scientific basis for precise pest and disease control, the following problems still exist:

[0005] There are many types of crop diseases and pests, and they are affected by various factors such as region, climate, and season. The model was only trained for some common diseases and pests, and the identification effect may be poor for rare or newly emerging diseases and pests. Furthermore, it is impossible to guarantee the real-time diagnosis of diseases and pests while maintaining a high diagnostic accuracy. Summary of the Invention

[0006] The purpose of this invention is to provide a smart agricultural product pest and disease diagnosis system and method based on the Internet of Things. By constructing a comprehensive pest and disease characteristic database and training a high-quality diagnostic model, the comprehensiveness and accuracy of the diagnostic system are improved, enabling real-time monitoring and rapid diagnosis of pests and diseases, thereby effectively improving prevention and control efficiency and solving the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] The IoT-based smart agricultural product pest and disease diagnosis system includes:

[0009] The image acquisition unit is used to deploy cameras and various sensors in farmland to capture real-time image data of crop growth environment and pests and diseases.

[0010] The feature extraction unit is used to establish a pest and disease feature database, construct a key feature recognition model based on the pest and disease feature database, process the collected pest and disease images, and extract key features from the pest and disease images based on the key feature recognition model.

[0011] The diagnostic decision unit is used to construct a disease and pest diagnosis model based on the extracted key features, identify and diagnose the types of diseases and pests affecting crops based on the disease and pest diagnosis model, determine the disease and pest diagnosis results, and send the real-time acquired image data and disease and pest diagnosis results to the remote interaction unit based on the Internet of Things.

[0012] The remote interaction unit is used to establish a data interaction channel with a remote terminal. Based on the data interaction channel, the terminal displays the diagnosis results of pests and diseases, their location, and prevention and control suggestions. At the same time, users can view historical diagnosis records and receive real-time alerts through the interactive interface.

[0013] Furthermore, the image acquisition unit includes:

[0014] The environmental parameter acquisition module is used to collect environmental parameters in farmland in real time based on various sensors, including temperature and humidity data, light intensity data, and soil moisture data;

[0015] The image acquisition module is used to capture real-time image data of crop growth status and pests and diseases based on the camera;

[0016] The data preprocessing module is used to clean and verify the integrity of the data collected from the environmental parameter acquisition module and the image acquisition module, and then transmit the preprocessed data to the feature extraction unit based on the packaged dataset.

[0017] Furthermore, the image acquisition module includes:

[0018] The image capture control module is used to control the camera to capture crop growth status and pest and disease image data in real time, and to use the growth status and pest and disease image data as target image data.

[0019] An image segmentation module is used to segment the target image data to obtain multiple image blocks corresponding to the target image data;

[0020] The grayscale extraction module is used to extract the grayscale values ​​of the pixel blocks contained in each image block;

[0021] The grayscale balance coefficient acquisition module is used to acquire the grayscale balance coefficient corresponding to each image block based on the grayscale values ​​of the pixel blocks contained in each image block; wherein, the grayscale balance coefficient is acquired by the following formula:

[0022]

[0023] Where K represents the grayscale balance coefficient corresponding to each image block; n represents the number of pixel blocks contained in each image block; S i S represents the grayscale value of the i-th pixel block; z S represents the center gray value corresponding to each image patch; h This represents the center gray value corresponding to the target image data;

[0024] The difference acquisition module is used to compare the grayscale balance coefficients of every two adjacent image blocks and obtain the difference corresponding to the grayscale balance coefficients.

[0025] The comparison module is used to compare the difference corresponding to the grayscale balance coefficient with a preset difference threshold.

[0026] The grayscale adjustment module is used to adjust the grayscale of the two adjacent image blocks that exceed the preset difference threshold when the difference between the grayscale balance coefficients of each two adjacent image blocks exceeds the preset difference threshold.

[0027] Furthermore, the grayscale adjustment module includes:

[0028] The image block extraction module is used to extract the two adjacent image blocks whose gray-scale balance coefficients exceed the preset difference threshold when the difference between the gray-scale balance coefficients of each two adjacent image blocks exceeds the preset difference threshold.

[0029] The image block marking module is used to mark two adjacent image blocks whose difference between the grayscale balance coefficients exceeds a preset difference threshold as the first image block and the second image block.

[0030] The first target grayscale value acquisition module is used to acquire the target grayscale value of each pixel block contained in the first image block using the grayscale balance coefficient of the first image block; wherein, the target grayscale value of each pixel block contained in the first image block is acquired by the following formula:

[0031]

[0032] Among them, S m01 S represents the target grayscale value of each pixel block contained in the first image block; 01 K represents the grayscale value of each pixel block contained in the first image block. 01K represents the grayscale balance coefficient corresponding to the first image block. 02 S represents the grayscale balance coefficient corresponding to the second image patch; 01z S represents the center gray value corresponding to the first image patch; 02z S represents the center gray value corresponding to the second image patch; h This represents the center gray value corresponding to the target image data; s 01 This represents the first digit parameter; the principle for determining the value of the first digit parameter is as follows:

[0033] when and When the signs of are the same, let s 01 =-1; when and When the signs of are different, let s 01 =1;

[0034] The first grayscale adjustment execution module is used to adjust each pixel block contained in the first image block according to the target grayscale value of each pixel block contained in the first image block;

[0035] The second target grayscale value acquisition module is used to acquire the target grayscale value of each pixel block contained in the second image block using the grayscale balance coefficient of the second image block; wherein, the target grayscale value of each pixel block contained in the second image block is acquired by the following formula:

[0036]

[0037] Among them, S m02 S represents the target grayscale value of each pixel block contained in the second image block; 02 K represents the grayscale value of each pixel block contained in the second image block. 01 K represents the grayscale balance coefficient corresponding to the first image block. 02 S represents the grayscale balance coefficient corresponding to the second image patch; 01z S represents the center gray value corresponding to the first image patch; 02z S represents the center gray value corresponding to the second image patch; h This represents the center gray value corresponding to the target image data; s 02 This represents the second bit parameter; meanwhile, the principles for determining the value of the first bit parameter are as follows:

[0038] when and When the signs of are the same, let s 02 =-1; when and When the signs of are different, let s 01 =1;

[0039] The second grayscale adjustment execution module is used to adjust each pixel block contained in the second image block according to the target grayscale value of each pixel block contained in the second image block.

[0040] Furthermore, the feature extraction unit includes:

[0041] The database construction module is used to crawl and store professional data on known pests and diseases based on the Internet of Things, including information on pest and disease images, feature descriptions, occurrence conditions, and control methods. Based on the crawled professional data, a pest and disease feature database is constructed.

[0042] The identification model building module is used to extract features based on professional data in the pest and disease feature database as sample data. The extracted features and corresponding pest and disease labels are used as training data and input into the preset model to build a key feature identification model.

[0043] The feature extraction module is used to extract key features of pests and diseases from the dataset based on the key feature recognition model, and to extract the environmental features of agricultural products in the dataset.

[0044] Furthermore, the recognition model construction module constructs a key feature recognition model, specifically as follows:

[0045] The professional data on the known pests and diseases are classified and integrated to obtain the processed classification data. Features are extracted from the processed classification data, and an initial feature set of the classification data is obtained based on the extraction results.

[0046] Retrieve key features related to pests and diseases from the initial feature set and integrate them into a key feature subset;

[0047] Obtain the time series data of each key feature in the key feature subset, determine the target feature corresponding to the time series data, and determine the first feature of each pest based on the target feature of each key feature;

[0048] Acquire the changes in the target characteristics of each disease and pest in crops, and determine the characteristic change rules of each disease and pest based on the changes.

[0049] Pests and diseases whose feature change rule similarity is greater than or equal to a preset threshold are identified as the same type of pests and diseases, and the second feature of any pest or disease in each type of pest and disease is identified as the final target feature of that type of pest and disease.

[0050] The final target features of this type of pest and disease are used as input samples for the model, while the feature information of each pest and disease is used as output samples to train the preset network model, so as to obtain the identification model of each type of pest and disease.

[0051] Furthermore, the feature extraction module also includes:

[0052] Based on the environmental characteristics of agricultural products, the extracted key features were screened to identify key features that are significantly affected by environmental characteristics and have a correlation with the occurrence of diseases and pests.

[0053] The weights of the key features in the pest and disease identification model are adjusted based on the degree to which the key features are affected by environmental features and the closeness of their correlation.

[0054] The pest and disease identification model is optimized based on the degree of influence of agricultural product environmental characteristics on key characteristics of pests and diseases and the weight of key characteristics in the pest and disease identification model.

[0055] Meanwhile, environmental adaptability analysis is performed on the extracted key features to determine the occurrence patterns and trends of pests and diseases under different environmental conditions. Furthermore, the occurrence trends of future pests and diseases are predicted based on the occurrence patterns and trends.

[0056] Furthermore, the diagnostic decision-making unit includes:

[0057] The feature diagnosis module is used to diagnose pests and diseases by extracting key features, and to determine the types of pests and diseases, the degree of impact on crops, and the scope of impact based on the diagnosis results.

[0058] The feature diagnosis module is also used to acquire abnormal data from the anomaly detection module and determine whether the abnormal data is rare or a newly emerging pest or disease characteristic.

[0059] The anomaly detection module is used to detect anomalies in crops that do not exhibit key characteristics of pests and diseases, and to identify abnormal situations that do not conform to normal growth patterns.

[0060] The pest and disease control management module is used to match the corresponding pest and disease control solutions based on the diagnostic results and abnormal situation data, evaluate the control results of the pest and disease control solutions, and generate feedback data based on the evaluation results.

[0061] Furthermore, the pest and disease control module matches the corresponding pest and disease control plan, specifically as follows:

[0062] Based on the diagnostic results and abnormal data, the severity of crop diseases and pests is determined, and a disease and pest control plan corresponding to the crop category is constructed based on the severity of crop diseases and pests.

[0063] The specific measures in the pest and disease control program include source control measures, pest and disease control process measures, and end-of-pipe treatment measures.

[0064] Analyze the pest and disease control plans and determine their priority.

[0065] Based on the specific measures of the pest and disease control plan, corresponding crop management personnel are assigned, and the crop management personnel implement the specific measures of the pest and disease control plan according to its priority.

[0066] This invention provides another technical solution: a smart agricultural product pest and disease diagnosis method based on the Internet of Things, comprising the following steps:

[0067] Step 1: Data Acquisition: Real-time image data of crop growth environment and pests and diseases are collected through a network of cameras and sensors deployed in farmland;

[0068] Step 2: Data Processing: Process the collected images of pests and diseases, and extract key features from the images based on the constructed key feature recognition model;

[0069] Step 3: Intelligent Diagnosis and Decision-Making: Real-time diagnosis of key feature inputs, and generation of pest and disease diagnosis reports and pest and disease control plans;

[0070] Step 4: Results Display and Early Warning: Display the pest and disease diagnosis report and pest and disease control plan to the user through the user interface, and send early warning information to remote terminals;

[0071] Step 5: Database Query and Prevention Suggestions: Users query the pest and disease feature database based on the obtained abnormal image data and diagnostic results to identify rare and newly emerging pests and diseases, and generate corresponding pest and disease prevention and control methods and suggestions.

[0072] Compared with the prior art, the beneficial effects of the present invention are:

[0073] The image acquisition unit captures real-time image data of crop growth environment and pests and diseases through cameras and sensors deployed in farmland, ensuring timely diagnosis. The feature extraction unit establishes a pest and disease feature database and constructs a key feature recognition model to achieve intelligent processing of pest and disease images, identifying and diagnosing various types of pests and diseases, including rare or newly emerging ones. As new data is continuously added, the diagnostic model is adaptively updated and optimized to improve its adaptability and accuracy. The remote interaction unit establishes a data interaction channel with remote terminals, enabling users to view pest and disease diagnosis results, location of occurrence, and control suggestions from anywhere via remote terminals. This improves user convenience, allows timely access to professional pest and disease control knowledge and assistance, enhances user response capabilities, reduces the cost of manual inspection and diagnosis, improves agricultural production efficiency, increases crop yield and quality, and reduces the impact of pests and diseases on crops. Attached Figure Description

[0074] Figure 1This is a block diagram of the IoT-based smart agricultural product pest and disease diagnosis system of the present invention;

[0075] Figure 2 This is a flowchart of the IoT-based smart agricultural product pest and disease diagnosis method of the present invention. Detailed Implementation

[0076] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0077] To address the challenges posed by the diverse range of crop diseases and pests, and their susceptibility to variations due to geographical location, climate, and season, the model was trained only on a subset of common diseases and pests. This may result in poor model recognition for rare or newly emerging diseases and pests, and it also fails to guarantee both real-time diagnostics and high diagnostic accuracy. Please refer to [link to relevant documentation]. Figure 1 This embodiment provides the following technical solution:

[0078] The IoT-based smart agricultural product pest and disease diagnosis system includes:

[0079] The image acquisition unit is used to deploy cameras and various sensors in farmland, such as temperature and humidity sensors, light sensors, and soil moisture sensors, to capture real-time image data of crop growth environment and pests and diseases, including:

[0080] The environmental parameter acquisition module is used to collect environmental parameters in farmland in real time based on various sensors, including temperature and humidity data, light intensity data, and soil moisture data;

[0081] The image acquisition module is used to capture real-time image data of crop growth status and pests and diseases based on the camera;

[0082] The data preprocessing module is used to clean and verify the integrity of the data collected from the environmental parameter acquisition module and the image acquisition module, and transmit the preprocessed data to the feature extraction unit based on the packaged dataset, thereby improving the accuracy and reliability of the data and providing a high-quality data foundation for subsequent data analysis and decision support.

[0083] The feature extraction unit is used to establish a pest and disease feature database, including images, feature descriptions, occurrence conditions, and control methods of known pests and diseases. Based on the pest and disease feature database, a key feature recognition model is constructed to provide training data and reference for the diagnostic model. At the same time, the collected pest and disease images are processed, including denoising, contrast enhancement, and image segmentation, to improve image quality and facilitate subsequent analysis. Based on the key feature recognition model, key features such as shape, color, and texture are extracted from the pest and disease images.

[0084] The diagnostic decision unit is used to construct a disease and pest diagnosis model based on the extracted key features, identify and diagnose the types of diseases and pests affecting crops based on the disease and pest diagnosis model, determine the disease and pest diagnosis results, and send the real-time acquired image data and disease and pest diagnosis results to the remote interaction unit based on the Internet of Things.

[0085] The remote interaction unit is used to establish a data interaction channel with a remote terminal. Based on the data interaction channel, the remote terminal displays the results of disease and pest diagnosis, the location of occurrence, and control suggestions. At the same time, users can view historical diagnosis records and receive real-time alerts through the interactive interface. The remote terminal can provide a rich knowledge base of disease and pest control and detailed help documents to help users better understand and deal with disease and pest problems.

[0086] In this embodiment, the image acquisition unit captures real-time image data of crop growth environment and pests and diseases through cameras and sensors deployed in the farmland, ensuring the timeliness of pest and disease diagnosis. The feature extraction unit establishes a pest and disease feature database and constructs a key feature recognition model to achieve intelligent processing of pest and disease images, identify and diagnose various types of pests and diseases, including rare or newly emerging ones. As new data is continuously added, the diagnostic model is adaptively updated and optimized to improve its adaptability and accuracy. The remote interaction unit constructs a data interaction channel with a remote terminal, enabling users to view pest and disease diagnosis results, occurrence locations, and prevention and control suggestions from anywhere via a remote terminal. This improves user convenience, allows timely access to professional pest and disease control knowledge and assistance, enhances user response capabilities, reduces the cost of manual inspection and diagnosis, improves agricultural production efficiency, increases crop yield and quality, and reduces the impact of pests and diseases on crops.

[0087] Specifically, the image acquisition module includes:

[0088] The image capture control module is used to control the camera to capture crop growth status and pest and disease image data in real time, and to use the growth status and pest and disease image data as target image data.

[0089] An image segmentation module is used to segment the target image data to obtain multiple image blocks corresponding to the target image data;

[0090] The grayscale extraction module is used to extract the grayscale values ​​of the pixel blocks contained in each image block;

[0091] The grayscale balance coefficient acquisition module is used to acquire the grayscale balance coefficient corresponding to each image block based on the grayscale values ​​of the pixel blocks contained in each image block; wherein, the grayscale balance coefficient is acquired by the following formula:

[0092]

[0093] Where K represents the grayscale balance coefficient corresponding to each image block; n represents the number of pixel blocks contained in each image block; S i S represents the grayscale value of the i-th pixel block; z S represents the center gray value corresponding to each image patch; h This represents the center gray value corresponding to the target image data;

[0094] The difference acquisition module is used to compare the grayscale balance coefficients of every two adjacent image blocks and obtain the difference corresponding to the grayscale balance coefficients.

[0095] The comparison module is used to compare the difference corresponding to the grayscale balance coefficient with a preset difference threshold.

[0096] The grayscale adjustment module is used to adjust the grayscale of the two adjacent image blocks that exceed the preset difference threshold when the difference between the grayscale balance coefficients of each two adjacent image blocks exceeds the preset difference threshold.

[0097] The technical effects of the above solution are as follows: The image capture control module can capture real-time image data of crop growth status and pests and diseases, providing a precise data foundation for subsequent image analysis and processing. The image segmentation module divides the target image data into multiple image blocks, facilitating independent processing and analysis of each image block and improving processing efficiency.

[0098] The grayscale value extraction module and the grayscale balance coefficient acquisition module work together to accurately extract the grayscale value of each image block and calculate the corresponding grayscale balance coefficient, providing a basis for subsequent grayscale adjustment. Through the combined action of the difference acquisition module, comparison module, and grayscale adjustment module, image blocks with unbalanced grayscale can be intelligently identified and adjusted accordingly, resulting in a more uniform grayscale across the entire image and improving image quality and analyzability. This technical solution enables clearer and more accurate image data on crop growth status and pests and diseases, thereby improving the accuracy of pest and disease identification and providing stronger technical support for agricultural production.

[0099] In summary, this technical solution effectively improves the quality and analysis efficiency of crop growth status and pest and disease image data through a series of steps, including real-time capture, precise segmentation, grayscale value extraction and balancing, and intelligent adjustment, providing more intelligent and precise technical support for agricultural production.

[0100] Specifically, the grayscale adjustment module includes:

[0101] The image block extraction module is used to extract the two adjacent image blocks whose gray-scale balance coefficients exceed the preset difference threshold when the difference between the gray-scale balance coefficients of each two adjacent image blocks exceeds the preset difference threshold.

[0102] The image block marking module is used to mark two adjacent image blocks whose difference between the grayscale balance coefficients exceeds a preset difference threshold as the first image block and the second image block.

[0103] The first target grayscale value acquisition module is used to acquire the target grayscale value of each pixel block contained in the first image block using the grayscale balance coefficient of the first image block; wherein, the target grayscale value of each pixel block contained in the first image block is acquired by the following formula:

[0104]

[0105] Among them, S m01 S represents the target grayscale value of each pixel block contained in the first image block; 01 K represents the grayscale value of each pixel block contained in the first image block. 01 K represents the grayscale balance coefficient corresponding to the first image block. 02 S represents the grayscale balance coefficient corresponding to the second image patch; 01z S represents the center gray value corresponding to the first image patch; 02z S represents the center gray value corresponding to the second image patch; h This represents the center gray value corresponding to the target image data; s 01 This represents the first digit parameter; the principle for determining the value of the first digit parameter is as follows:

[0106] when and When the signs of are the same, let s 01 =-1; when and When the signs of are different, let s 01 =1;

[0107] The first grayscale adjustment execution module is used to adjust each pixel block contained in the first image block according to the target grayscale value of each pixel block contained in the first image block;

[0108] The second target grayscale value acquisition module is used to acquire the target grayscale value of each pixel block contained in the second image block using the grayscale balance coefficient of the second image block; wherein, the target grayscale value of each pixel block contained in the second image block is acquired by the following formula:

[0109]

[0110] Among them, S m02 S represents the target grayscale value of each pixel block contained in the second image block; 02 K represents the grayscale value of each pixel block contained in the second image block. 01 K represents the grayscale balance coefficient corresponding to the first image block. 02 S represents the grayscale balance coefficient corresponding to the second image patch; 01z S represents the center gray value corresponding to the first image patch; 02z S represents the center gray value corresponding to the second image patch; h This represents the center gray value corresponding to the target image data; s 02 This represents the second bit parameter; meanwhile, the principles for determining the value of the first bit parameter are as follows:

[0111] when and When the signs of are the same, let s 02 =-1; when and When the signs of are different, let s 01 =1;

[0112] The second grayscale adjustment execution module is used to adjust each pixel block contained in the second image block according to the target grayscale value of each pixel block contained in the second image block.

[0113] The technical effect of the above solution is as follows: the image block extraction module can accurately identify adjacent image blocks with gray-level imbalance, providing precise positioning for subsequent gray-level adjustment. Using the first target gray-level value acquisition module and the second target gray-level value acquisition module, the target gray-level value of each pixel block in two adjacent image blocks can be calculated respectively, achieving precise adjustment of gray-level imbalance.

[0114] When calculating the target grayscale value, a first-digit parameter and a second-digit parameter are introduced. The selection of these parameters considers the grayscale balance coefficient and the difference in center grayscale value, making grayscale adjustment more intelligent and accurate. Through the grayscale adjustment execution module, each pixel block in the image can be adjusted according to the calculated target grayscale value, resulting in a more uniform grayscale across the entire image, improving image quality and analyzability. The grayscale-adjusted image is clearer, and the characteristics of pests and diseases are more prominent, helping to improve the accuracy of pest and disease identification and providing stronger technical support for agricultural production. Through modular design, each module undertakes a specific task, making the entire grayscale adjustment process more efficient and orderly.

[0115] In summary, this technical solution provides more intelligent and precise technical support for image analysis and processing in agricultural production through a series of advantages, including precise grayscale adjustment, intelligent parameter selection, improved image quality, enhanced pest and disease identification capabilities, and increased processing efficiency.

[0116] In this embodiment, the feature extraction unit includes:

[0117] The database construction module is used to crawl and store professional data on known pests and diseases based on the Internet of Things, including information on pest and disease images, feature descriptions, occurrence conditions, and control methods. Based on the crawled professional data, a pest and disease feature database is constructed.

[0118] The identification model building module is used to extract features based on professional data in the pest and disease feature database as sample data. The extracted features and corresponding pest and disease labels are used as training data and input into the preset model to build a key feature identification model.

[0119] The feature extraction module is used to extract key features of pests and diseases from the dataset based on a key feature recognition model, and to extract environmental features of agricultural products from the dataset. It also includes:

[0120] Based on the environmental characteristics of agricultural products, the extracted key features were screened to identify key features that are significantly affected by environmental characteristics and have a correlation with the occurrence of diseases and pests.

[0121] The weights of the key features in the pest and disease identification model are adjusted according to the degree to which the key features are affected by environmental features and the tightness of their correlation, thereby realizing the intelligence and adaptability of the model.

[0122] The pest and disease identification model is optimized based on the degree of influence of agricultural product environmental characteristics on key characteristics of pests and diseases and the weight of key characteristics in the pest and disease identification model.

[0123] Meanwhile, environmental adaptability analysis is performed on the extracted key features to determine the occurrence patterns and trends of pests and diseases under different environmental conditions. Furthermore, the occurrence trends of future pests and diseases are predicted based on the occurrence patterns and trends.

[0124] In this embodiment, extensive and in-depth professional data on known pests and diseases are crawled and stored using IoT technology, ensuring the accuracy and richness of the pest and disease feature database. This lays a solid foundation for subsequent feature extraction and model training. The identification model building module not only improves the accuracy of the model but also enables it to identify various pests and diseases, including rare or newly emerging ones. While extracting key features of pests and diseases, the influence of agricultural product environmental characteristics on the occurrence of pests and diseases is also considered. Based on the occurrence patterns and trends, the future occurrence of pests and diseases is predicted, providing farmers with timely prevention and control suggestions, reducing the impact of pests and diseases on crops. By comprehensively considering agricultural product environmental characteristics and key pest and disease characteristics, pests and diseases are diagnosed more accurately, reducing false alarms and missed alarms, and improving diagnostic efficiency.

[0125] In this embodiment, the recognition model construction module constructs a key feature recognition model, specifically as follows:

[0126] The professional data on the known pests and diseases are classified and integrated to obtain the processed classification data. Features are extracted from the processed classification data, and an initial feature set of the classification data is obtained based on the extraction results.

[0127] Retrieve key features related to pests and diseases from the initial feature set and integrate them into a key feature subset;

[0128] Obtain the time series data of each key feature in the key feature subset, determine the target feature corresponding to the time series data, and determine the first feature of each pest based on the target feature of each key feature;

[0129] Acquire the changes in the target characteristics of each disease and pest in crops, and determine the characteristic change rules of each disease and pest based on the changes.

[0130] Identifying pests and diseases with a similarity of feature change rules greater than or equal to a preset threshold as the same type of pests and diseases helps to simplify the complexity of the model. The second feature of any pest or disease in each type of pest and disease is identified as the final target feature of that type of pest or disease.

[0131] The final target features of this type of pest and disease are used as input samples for the model, while the feature information of each pest and disease is used as output samples to train the preset network model, so as to obtain the identification model of each type of pest and disease.

[0132] In this embodiment, classification processing can more effectively extract key features related to pests and diseases, providing a high-quality data foundation for subsequent model construction. Key features related to pests and diseases are retrieved from the initial feature set and integrated into a key feature subset, which not only improves the efficiency of feature extraction but also ensures the comprehensiveness and representativeness of key features. The feature change rules determined by analyzing the changes of pests and diseases in crops can more realistically reflect the biological characteristics and occurrence patterns of pests and diseases, ensuring that the identification model for each type of pest and disease can more accurately identify different manifestations of the same type of pest and disease, thus improving the model's recognition accuracy. Furthermore, since the model is trained based on data from multiple types of pests and diseases, it also has strong generalization ability and can identify new or unknown pests and diseases.

[0133] In this embodiment, the diagnostic decision unit includes:

[0134] The feature diagnosis module is used to diagnose pests and diseases by extracting key features, and to determine the types of pests and diseases, the degree of impact on crops, and the scope of impact based on the diagnosis results.

[0135] The feature diagnosis module is also used to acquire abnormal data from the anomaly detection module and determine whether the abnormal data is rare or a newly emerging pest or disease characteristic.

[0136] The anomaly detection module is used to detect anomalies in crops that do not exhibit key characteristics of pests and diseases, and to identify abnormal situations that do not conform to normal growth patterns, including sudden changes in crop color, slowed growth rate, and wilting leaves.

[0137] The pest and disease control management module is used to match the corresponding pest and disease control solutions based on the diagnostic results and abnormal situation data, evaluate the control results of the pest and disease control solutions, and generate feedback data based on the evaluation results.

[0138] In this embodiment, the pest and disease control module matches the corresponding pest and disease control plan, specifically as follows:

[0139] Based on the diagnostic results and abnormal data, the severity of crop diseases and pests is determined, and a disease and pest control plan corresponding to the crop category is constructed based on the severity of crop diseases and pests.

[0140] The specific measures in the pest and disease control program include source control measures, pest and disease control process measures, and end-of-pipe treatment measures.

[0141] Analyze the pest and disease control plans and determine their priority.

[0142] Based on the specific measures of the pest and disease control plan, corresponding crop management personnel are assigned. The crop management personnel implement the specific measures of the pest and disease control plan according to its priority. By assigning crop management personnel and implementing specific control measures, efficient coordination between various links can be achieved, which not only improves control efficiency but also ensures the effective implementation of control measures.

[0143] In this embodiment, the feature diagnosis module can accurately extract key features of pests and diseases and diagnose them based on these features. It can not only quickly determine the type of pests and diseases, but also assess their impact on crops and their extent, and identify rare or newly emerging pest and disease characteristics, providing the possibility for timely response. The anomaly detection module monitors the growth status of crops and can promptly detect abnormal situations that do not conform to the normal growth pattern, providing timely warnings to prevent the spread of pests and diseases and the loss of crop yield. It not only considers the type and level of pests and diseases, but also combines the specific conditions of crops and their growth environment. By assigning crop management personnel and implementing specific control measures, it achieves scientific management of pests and diseases, ensuring the healthy growth of crops and the stable increase in yield.

[0144] To better demonstrate the implementation of the IoT-based smart agricultural product pest and disease diagnosis system, please refer to [link / reference needed]. Figure 2 This invention provides another method for diagnosing pests and diseases in agricultural products based on the Internet of Things, comprising the following steps:

[0145] Step 1: Data Acquisition: Real-time image data of crop growth environment and pests and diseases are collected through a network of cameras and sensors deployed in farmland;

[0146] Step 2: Data Processing: Process the collected images of pests and diseases, and extract key features from the images based on the constructed key feature recognition model;

[0147] Step 3: Intelligent Diagnosis and Decision-Making: Real-time diagnosis of key feature inputs, and generation of pest and disease diagnosis reports and pest and disease control plans;

[0148] Step 4: Results Display and Early Warning: Display the pest and disease diagnosis report and pest and disease control plan to the user through the user interface, and send early warning information to remote terminals;

[0149] Step 5: Database Query and Prevention Suggestions: Users query the pest and disease feature database based on the obtained abnormal image data and diagnostic results to identify rare and newly emerging pests and diseases, and generate corresponding pest and disease prevention and control methods and suggestions.

[0150] In this embodiment, by integrating real-time data acquisition from cameras and sensor networks, image features are processed and identified to enable accurate and rapid diagnosis and decision-making regarding pests and diseases. Simultaneously, user-friendly database query and prevention / control suggestion functions are provided, significantly improving the efficiency and accuracy of crop pest and disease management, reducing production costs, protecting the ecological environment, and enhancing farmers' ability to cope with pests and diseases, thus providing strong support for the sustainable development of agricultural production.

[0151] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A smart agricultural product pest and disease diagnosis system based on the Internet of Things, characterized in that: include: The image acquisition unit is used to deploy cameras and various sensors in farmland to capture real-time image data of crop growth environment and pests and diseases. The feature extraction unit is used to establish a pest and disease feature database, construct a key feature recognition model based on the pest and disease feature database, process the collected pest and disease images, and extract key features from the pest and disease images based on the key feature recognition model. The diagnostic decision unit is used to build a disease and pest diagnosis model based on the extracted key features, identify and diagnose the types of diseases and pests that crops are suffering from based on the disease and pest diagnosis model, determine the disease and pest diagnosis results, and send the real-time acquired image data and disease and pest diagnosis results to the remote interaction unit based on the Internet of Things. The remote interaction unit is used to build a data interaction channel with a remote terminal, and based on the data interaction channel, it displays the disease and pest diagnosis results, occurrence location and control suggestions on the remote terminal. Meanwhile, users can view historical diagnostic records and receive real-time alerts through the interactive interface; The feature extraction unit includes: The database construction module is used to crawl and store professional data on known pests and diseases based on the Internet of Things, including information on pest and disease images, feature descriptions, occurrence conditions, and control methods. Based on the crawled professional data, a pest and disease feature database is constructed. The identification model building module is used to extract features based on professional data in the pest and disease feature database as sample data. The extracted features and corresponding pest and disease labels are used as training data and input into the preset model to build a key feature identification model. The feature extraction module is used to extract key features of pests and diseases from the dataset based on the key feature recognition model, and to extract environmental features of agricultural products from the dataset; the dataset is a preprocessed dataset generated by packaging data collected from the environmental parameter acquisition module and the image acquisition module. The identification model construction module constructs a key feature identification model, specifically as follows: The professional data on the known pests and diseases are classified and integrated to obtain the processed classification data. Features are extracted from the processed classification data, and an initial feature set of the classification data is obtained based on the extraction results. Key features related to pests and diseases are retrieved from the initial feature set and integrated into a subset of key features; Obtain the time series data of each key feature in the key feature subset, determine the target feature corresponding to the time series data, and determine the first feature of each pest based on the target feature of each key feature; Acquire the changes in the target characteristics of each disease and pest in crops, and determine the characteristic change rules of each disease and pest based on the changes. Pests and diseases whose feature change rule similarity is greater than or equal to a preset threshold are identified as the same type of pests and diseases, and the second feature of any pest or disease in each type of pest and disease is identified as the final target feature of that type of pest and disease. The final target features of this type of pest and disease are used as input samples for the model, and the feature information of each pest and disease is used as output samples for the model to train the preset network model, so as to obtain the key feature recognition model for each type of pest and disease. The feature extraction module further includes: Based on the environmental characteristics of agricultural products, the extracted key features were screened to identify key features that are significantly affected by environmental characteristics and have a correlation with the occurrence of diseases and pests. The weights of the key features in the pest and disease identification model are adjusted based on the degree to which the key features are affected by environmental features and the closeness of their correlation. The pest and disease identification model is optimized based on the degree of influence of agricultural product environmental characteristics on key characteristics of pests and diseases and the weight of key characteristics in the pest and disease identification model. Meanwhile, environmental adaptability analysis is performed on the extracted key features to determine the occurrence patterns and trends of pests and diseases under different environmental conditions. Furthermore, the occurrence trends of future pests and diseases are predicted based on the occurrence patterns and trends.

2. The IoT-based smart agricultural product pest and disease diagnosis system as described in claim 1, characterized in that: The image acquisition unit includes: The environmental parameter acquisition module is used to collect environmental parameters in farmland in real time based on various sensors, including temperature and humidity data, light intensity data, and soil moisture data; The image acquisition module is used to capture real-time image data of crop growth status and pests and diseases based on the camera; The data preprocessing module is used to clean and verify the integrity of the data collected from the environmental parameter acquisition module and the image acquisition module, and then transmit the preprocessed data to the feature extraction unit based on the packaged dataset.

3. The IoT-based smart agricultural product pest and disease diagnosis system as described in claim 2, characterized in that: The image acquisition module includes: The image capture control module is used to control the camera to capture crop growth status and pest and disease image data in real time, and to use the growth status and pest and disease image data as target image data. An image segmentation module is used to segment the target image data to obtain multiple image blocks corresponding to the target image data; The grayscale extraction module is used to extract the grayscale values ​​of the pixel blocks contained in each image block; The grayscale balance coefficient acquisition module is used to acquire the grayscale balance coefficient corresponding to each image block based on the grayscale values ​​of the pixel blocks contained in each image block; wherein, the grayscale balance coefficient is acquired by the following formula: Where K represents the grayscale balance coefficient corresponding to each image block; n represents the number of pixel blocks contained in each image block; S i S represents the grayscale value of the i-th pixel block; z S represents the center gray value corresponding to each image patch; h This represents the center gray value corresponding to the target image data; The difference acquisition module is used to compare the grayscale balance coefficients of every two adjacent image blocks and obtain the difference corresponding to the grayscale balance coefficients. The comparison module is used to compare the difference corresponding to the grayscale balance coefficient with a preset difference threshold. The grayscale adjustment module is used to adjust the grayscale of two adjacent image blocks that exceed the preset difference threshold when the difference between the grayscale balance coefficients of each two adjacent image blocks exceeds the preset difference threshold.

4. The IoT-based smart agricultural product pest and disease diagnosis system as described in claim 3, characterized in that: The grayscale adjustment module includes: The image block extraction module is used to extract the two adjacent image blocks whose gray-scale balance coefficients exceed the preset difference threshold when the difference between the gray-scale balance coefficients of each two adjacent image blocks exceeds the preset difference threshold. The image block marking module is used to mark two adjacent image blocks whose difference between the grayscale balance coefficients exceeds a preset difference threshold as the first image block and the second image block. The first target grayscale value acquisition module is used to acquire the target grayscale value of each pixel block contained in the first image block using the grayscale balance coefficient of the first image block; wherein, the target grayscale value of each pixel block contained in the first image block is acquired by the following formula: Among them, S m01 S represents the target grayscale value of each pixel block contained in the first image block; 01 K represents the grayscale value of each pixel block contained in the first image block. 01 K represents the grayscale balance coefficient corresponding to the first image block. 02 S represents the grayscale balance coefficient corresponding to the second image patch; 01z S represents the center gray value corresponding to the first image patch; 02z S represents the center gray value corresponding to the second image patch; h This represents the center gray value corresponding to the target image data; s 01 This represents the first digit parameter; the principle for determining the value of the first digit parameter is as follows: when and When the signs of are the same, let s 01 =-1; when and When the signs of are different, let s 01 =1; The first grayscale adjustment execution module is used to adjust each pixel block contained in the first image block according to the target grayscale value of each pixel block contained in the first image block; The second target grayscale value acquisition module is used to acquire the target grayscale value of each pixel block contained in the second image block using the grayscale balance coefficient of the second image block; wherein, the target grayscale value of each pixel block contained in the second image block is acquired by the following formula: Among them, S m02 S represents the target grayscale value of each pixel block contained in the second image block; 02 K represents the grayscale value of each pixel block contained in the second image block. 01 K represents the grayscale balance coefficient corresponding to the first image block. 02 S represents the grayscale balance coefficient corresponding to the second image patch; 01z S represents the center gray value corresponding to the first image patch; 02z S represents the center gray value corresponding to the second image patch; h This represents the center gray value corresponding to the target image data; s 02 This represents the second bit parameter; meanwhile, the principles for determining the value of the first bit parameter are as follows: when and When the signs of are the same, let s 02 =-1; when and When the signs of are different, let s 02 =1; The second grayscale adjustment execution module is used to adjust each pixel block contained in the second image block according to the target grayscale value of each pixel block contained in the second image block.

5. The IoT-based smart agricultural product pest and disease diagnosis system as described in claim 4, characterized in that: The diagnostic decision-making unit includes: The feature diagnosis module is used to diagnose pests and diseases by extracting key features, and to determine the types of pests and diseases, the degree of impact on crops, and the scope of impact based on the diagnosis results. The feature diagnosis module is also used to acquire abnormal data from the anomaly detection module and determine whether the abnormal data is rare or a newly emerging pest or disease characteristic. The anomaly detection module is used to detect anomalies in crops that do not exhibit key characteristics of pests and diseases, and to identify abnormal situations that do not conform to normal growth patterns. The pest and disease control management module is used to match the corresponding pest and disease control solutions based on the diagnostic results and abnormal situation data, evaluate the control results of the pest and disease control solutions, and generate feedback data based on the evaluation results.

6. The IoT-based smart agricultural product pest and disease diagnosis system as described in claim 5, characterized in that: The pest and disease control module matches the corresponding pest and disease control plan with the treatment plan, specifically: Based on the diagnostic results and abnormal data, the severity of crop diseases and pests is determined, and a disease and pest control plan corresponding to the crop category is constructed based on the severity of crop diseases and pests. The specific measures in the pest and disease control program include source control measures, pest and disease control process measures, and end-of-pipe treatment measures. Analyze the pest and disease control plans and determine their priority. Based on the specific measures of the pest and disease control plan, corresponding crop management personnel are assigned, and the crop management personnel implement the specific measures of the pest and disease control plan according to its priority.

7. A smart agricultural product pest and disease diagnosis method based on the Internet of Things, applied in any one of claims 1-6, characterized in that: Includes the following steps: Step 1: Data Acquisition: Real-time image data of crop growth environment and pests and diseases are collected through a network of cameras and sensors deployed in farmland; Step 2: Data Processing: Process the collected images of pests and diseases, and extract key features from the images based on the constructed key feature recognition model; Step 3: Intelligent Diagnosis and Decision-Making: Real-time diagnosis of key feature inputs, and generation of pest and disease diagnosis reports and pest and disease control plans; Step 4: Results Display and Early Warning: Display the pest and disease diagnosis report and pest and disease control plan to the user through the user interface, and send early warning information to remote terminals; Step 5: Database Query and Prevention Suggestions: Users query the pest and disease feature database based on the obtained abnormal image data and diagnostic results to identify rare and newly emerging pests and diseases, and generate corresponding pest and disease prevention and control methods and suggestions.

Images