Intelligent agricultural information big data intelligent acquisition management system and method

Abstract

The invention relates to the technical field of information, in particular to an intelligent agricultural information big data intelligent acquisition management system and method. The method comprises the following steps: acquiring farm sensing data, and performing crop growth analysis according to the farm sensing data so as to obtain crop growth data; performing crop health state feature extraction and crop physiological feature extraction according to the crop growth data so as to obtain crop health state data and crop physiological data; performing disease and pest detection according to the crop health state data and the crop physiological data to obtain crop disease and pest data; the method comprises the following steps: acquiring farm remote sensing data, performing crop multispectral analysis according to the farm remote sensing data so as to obtain crop multispectral data, and performing crop moisture state evaluation according to the crop multispectral data so as to obtain crop moisture state data. The pesticide spraying strategy is optimized based on the information technology, and the agricultural production management efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a smart agricultural information big data intelligent collection management system and method. Background Art

[0002] Traditional crop information collection systems often rely on traditional databases and data processing algorithms, which are inefficient when processing large-scale, multi-dimensional data. As the amount of data continues to increase, it is difficult for traditional methods to quickly process and analyze data from different sources (such as remote sensing images, soil moisture, climate data, etc.), resulting in delays in information processing and affecting the execution of real-time decisions. For complex data associations and trend predictions, the accuracy and applicability of traditional algorithms are also relatively limited, and they cannot effectively support intelligent and automated agricultural management. Traditional crop information collection and management systems usually rely on static data collection and processing and lack effective dynamic feedback mechanisms. During the growth of crops, the growth status of crops will change due to factors such as climate change, soil conditions, or pests and diseases. Traditional systems are often unable to dynamically adjust management strategies such as irrigation, fertilization, and pesticide spraying according to real-time changes. This makes it difficult to refine and intelligentize the agricultural production management process, which can easily lead to waste of resources or poor crop growth. Summary of the invention

[0003] Based on this, it is necessary for the present invention to provide a smart agricultural information big data intelligent collection management system and method to solve at least one of the above technical problems.

[0004] To achieve the above purpose, a method for intelligent collection and management of smart agricultural information big data includes the following steps:

[0005] Step S1: acquiring farm sensor data, and performing crop growth analysis based on the farm sensor data, thereby obtaining crop growth data;

[0006] Step S2: extracting crop health status characteristics and crop physiological characteristics according to crop growth data, thereby obtaining crop health status data and crop physiological data; performing pest and disease detection according to the crop health status data and crop physiological data, thereby obtaining crop pest and disease data;

[0007] Step S3: acquiring farm remote sensing data, and performing crop multispectral analysis based on the farm remote sensing data, thereby obtaining crop multispectral data, and performing crop moisture status assessment based on the crop multispectral data, thereby obtaining crop moisture status data;

[0008] Step S4: performing correlation analysis on the crop moisture status data according to the crop pest data, thereby obtaining crop pest-moisture status data; visualizing the pest-affected area according to the crop pest-moisture status data, thereby obtaining pest-affected area visualization data;

[0009] Step S5: Perform intelligent management of pesticide spraying according to the visualized data of the area affected by pests and diseases, thereby obtaining intelligent management data of pesticide spraying.

[0010] The present invention can accurately monitor the growth status of crops by acquiring farm sensor data and performing crop growth analysis, and provide reliable data support for subsequent health status and physiological feature extraction. This data-driven analysis method can identify potential problems in the crop growth process, give early warnings, and avoid damage to crop health. Furthermore, through the pest and disease detection function, pests encountered by crops can be discovered in time, reducing the errors of traditional pest and disease monitoring, and improving the accuracy and timeliness of pest and disease control. The system combines remote sensing data and uses multispectral analysis to evaluate the moisture status of crops. This method can more accurately reflect the moisture status of crops, and is more timely and comprehensive than traditional soil moisture detection. Through the correlation analysis of pest and disease data and moisture status data, the system can reveal the mutual influence between pests and diseases and moisture status, and optimize irrigation and prevention strategies in agricultural production management. In addition, the pest and disease impact area visualization function can clearly display the areas affected by pests and diseases, providing a decision-making basis for precise pesticide application and irrigation. Finally, by optimizing the intelligent management of pesticide spraying, the spraying amount can be dynamically adjusted according to the actual pest and disease affected area and moisture status to avoid resource waste, while improving the production efficiency of crops and the efficiency of pesticide utilization and reducing the impact on the environment. The organic combination of this series of steps effectively improves the level of intelligent agricultural management, ensures the rational use of agricultural resources, and improves the health and yield of crop growth, achieving the goal of refined and automated agricultural management.

[0011] Optionally, step S1 specifically includes:

[0012] Step S11: acquiring farm sensor data, and performing nutrient sensor feature extraction based on the farm sensor data, thereby obtaining nutrient sensor data;

[0013] Step S12: Counting soil nutrient element concentrations according to the nutrient sensing data, thereby obtaining soil nutrient element concentration data;

[0014] Step S13: Acquire crop nutrient requirement data and crop growth stage data;

[0015] Step S14: constructing a crop nutrient requirement model according to the crop nutrient requirement data and the crop growth stage data, thereby obtaining a crop nutrient requirement model;

[0016] Step S15: identifying crop nutrient ratio imbalance based on the soil nutrient element concentration data according to the crop nutrient requirement model, thereby obtaining crop nutrient ratio imbalance data;

[0017] Step S16: extracting chlorophyll meter features according to farm sensor data, thereby obtaining chlorophyll meter data;

[0018] Step S17: Analyze the chlorophyll content of crops on the chlorophyll meter data to obtain the chlorophyll content of crops;

[0019] Step S18: Perform crop growth integration according to the crop chlorophyll content and crop nutrient ratio imbalance data to obtain crop growth data.

[0020] The present invention provides accurate data support for the nutrient status of the soil by acquiring farm sensor data and performing nutrient sensor feature extraction, which can help farm managers better understand the nutrients in the soil and their changes, thereby providing a basis for rational fertilization and optimization of agricultural resources. Based on nutrient sensor data, the system can accurately count the concentration of nutrient elements in the soil, timely discover the situation of insufficient or excessive soil nutrients, and avoid the inefficient monitoring of soil nutrient status and data lag problems in traditional methods. Combined with the nutrient demand data and growth stage data of crops, the system can construct a crop nutrient demand model to provide scientific guidance for the nutrient demand of crops at different growth stages. The establishment of this model enables farm managers to fertilize more accurately according to the actual needs of crops, avoiding nutrient waste and insufficient supply. Through the analysis of soil nutrient element concentration data, the system can identify the imbalance of nutrient ratios of crops, timely warn and provide adjustment plans, thereby effectively avoiding crop growth problems caused by nutrient imbalance. Furthermore, by extracting chlorophyll meter features and analyzing chlorophyll content, the system can directly reflect the nutritional status and growth health of crops, helping farm managers to grasp the photosynthesis efficiency and growth performance of crops in real time. Combined with nutrient imbalance data, the system can comprehensively integrate the growth of crops, generate accurate growth data, and guide farms to take targeted management measures. By dynamically adjusting agricultural management strategies such as fertilization and irrigation, the system can achieve the optimal allocation of agricultural resources, avoid resource waste, and improve crop yield and quality. Ultimately, this intelligent management method based on big data and real-time feedback has greatly improved the level of refinement of agricultural production, promoted sustainable agricultural development, and effectively improved the growth quality and economic benefits of crops.

[0021] Optionally, step S17 is specifically:

[0022] Step S171: performing reflected light intensity statistics on the chlorophyll meter data to obtain reflected light intensity data;

[0023] Step S172: performing band division according to the reflected light intensity data, thereby obtaining red light band data and near infrared light band data;

[0024] Step S173: Calculating the reflectivity of the red light band data, thereby obtaining the red light band reflectivity data;

[0025] Step S174: calculating the reflectivity of the near-infrared light band data, thereby obtaining the near-infrared light band reflectivity data;

[0026] Step S175: evaluating the chlorophyll content of crops according to the red light band reflectance data and the near infrared light band reflectance data, thereby obtaining the chlorophyll content data of crops.

[0027] The present invention can accurately obtain the optical reflection characteristics of crop leaves by counting the reflected light intensity of chlorophyll meter data, which provides basic data for analyzing the growth status of crops. Using these reflected light intensity data, the system can further divide the bands and obtain specific data of the red light band and the near-infrared light band, which enables the information of different bands to be effectively separated, thereby improving the accuracy of subsequent analysis. By calculating the reflectivity of the red light band data and the near-infrared light band data, the system can accurately evaluate the reflectivity of crop leaves, which is crucial for judging the chlorophyll content of crops. The calculation of reflectivity can help distinguish the growth health of different crops and effectively evaluate their photosynthesis capacity. Finally, combined with the reflectivity data of the red light band and the near-infrared light band, the system can accurately evaluate the chlorophyll content of crops and derive the health status and growth ability of crops. The real-time and accuracy of the whole process enable agricultural managers to promptly discover crop growth problems and take corresponding management measures, such as adjusting irrigation, fertilization or pest control strategies, thereby realizing dynamic regulation and precise management of crop growth. These steps effectively avoid the delays and inaccuracies in traditional agricultural management methods, promote the process of intelligent and automated agriculture, increase crop yield and quality, and reduce resource waste and environmental burden.

[0028] Optionally, step S2 specifically includes:

[0029] Step S21: extracting crop health status characteristics and crop physiological characteristics according to crop growth data, thereby obtaining crop health status data and crop physiological data;

[0030] Step S22: performing pest feeding trace analysis on the crop health status data to obtain pest feeding trace data;

[0031] Step S23: performing wormhole analysis on crop physiological data to obtain wormhole data;

[0032] Step S24: Integrate the characteristics of crop pests and diseases based on the pest feeding trace data and the wormhole data, so as to obtain crop pest and disease data.

[0033] By extracting the health status characteristics of crop growth data, the system can understand the growth status of crops in detail and extract relevant physiological characteristics, which lays the foundation for further analyzing the health status and growth potential of crops. Based on these health and physiological data, the system can perform pest trace analysis, accurately identify the specific parts of the crop surface attacked by pests, and further provide detailed information on the occurrence of pests. Through wormhole analysis, the system can also reveal whether the inside of the crop is invaded by pests, which is crucial for a comprehensive assessment of the impact of pests and diseases. Combining pest trace data with wormhole data, the system can integrate a full range of crop pest and disease characteristics, thereby more accurately identifying the type and severity of pests and diseases. This integrated analysis not only improves the accuracy of pest and disease monitoring, but also can timely reflect the risks faced by crops during growth, so that farm managers can quickly take effective prevention and control measures, thereby reducing crop losses and increasing yields. In addition, intelligent data processing and real-time feedback mechanisms can also dynamically adjust strategies in agricultural management, such as precise application of pesticides and adjustment of crop growth environment, avoid waste of resources, and promote the refinement and intelligence of crop management.

[0034] Optionally, step S22 is specifically:

[0035] Step S221: collecting images of crop health status data to obtain crop health status images;

[0036] Step S222: performing grayscale conversion according to the crop health status image, thereby obtaining a crop health status grayscale image;

[0037] Step S223: extracting leaf texture features according to the grayscale image of the crop health status, thereby obtaining leaf texture data;

[0038] Step S224: performing roughness statistics on the blade texture data to obtain high-roughness blade texture data;

[0039] Step S225: dividing the crop health grayscale image into regions according to the high-roughness leaf texture data, thereby obtaining high-roughness leaf texture region data;

[0040] Step S226: performing linear food mark recognition according to the leaf texture data, thereby obtaining linear food mark data;

[0041] Step S227: performing an insect feeding trace intersection operation based on the linear feeding trace data and the high-roughness leaf texture area data to obtain the insect feeding trace data.

[0042] The present invention obtains crop health status images through image acquisition, and the system can obtain visual information of crops in real time, providing raw data for subsequent analysis. By using grayscale conversion technology, crop health status images can be converted into grayscale images, which simplifies the image processing process, highlights the key features of the image, and provides a clear data basis for further analysis. Based on the grayscale image, extracting leaf texture features can reveal the microstructure of crop leaves and provide more detailed information for judging the health status of crops, especially when crop leaves are invaded by pests and diseases, the changes in texture features are particularly obvious. By counting the roughness of leaf texture data, those leaf texture areas with high roughness can be identified, which usually indicate the parts that are more seriously damaged, and further determine the specific location of the occurrence of pests. Combined with high-roughness leaf texture data, the grayscale image of crop health status is divided into regions, which helps to locate the areas where pests and diseases are most concentrated, thereby providing a clear target for subsequent prevention and control. Linear food marks are identified based on leaf texture data, which can accurately identify pest food marks on crop leaves, help quickly locate the types and distribution of pests, and then judge their specific impact on crop health. By performing intersection operations on linear food trace data and high-roughness leaf texture area data, pest traces can be more accurately identified, thereby obtaining more accurate pest food trace data. The combination of this series of steps not only improves the accuracy and real-time nature of pest monitoring, but also dynamically monitors crop health status through image analysis, ensuring that agricultural management decisions can respond to pest threats in a timely and effective manner, ultimately achieving the goal of optimizing resource allocation and reducing crop losses.

[0043] Optionally, step S226 is specifically:

[0044] Gray-level co-occurrence matrix calculation is performed according to leaf texture data, thereby obtaining gray-level co-occurrence matrix data;

[0045] Perform entropy statistics on the gray-level co-occurrence matrix data to obtain high-entropy gray-level co-occurrence matrix data;

[0046] Energy calculation is performed according to the gray-level co-occurrence matrix data, thereby obtaining gray-level co-occurrence matrix energy data;

[0047] Perform energy statistics on the gray-level co-occurrence matrix energy data to obtain low-energy gray-level co-occurrence matrix data;

[0048] The food trace data is obtained by performing intersection operation on the low-energy gray-level co-occurrence matrix data and the high-entropy gray-level co-occurrence matrix data;

[0049] The food trace data is subjected to elongated shape recognition to obtain linear food trace data.

[0050] The present invention calculates the grayscale co-occurrence matrix of leaf texture data, and the system can extract spatial texture information from the image, further revealing the microstructural characteristics of crop leaves. The grayscale co-occurrence matrix can provide the spatial relationship between the grayscale of pixels in the image, reflecting the health of the leaves and potential signs of pests and diseases. Entropy statistics based on grayscale co-occurrence matrix data can measure the complexity and information content of the image. The higher the entropy value, the more complex the texture in the image, and also indicates the presence of pests and diseases or the stress response of crops. By extracting high-entropy grayscale co-occurrence matrix data, it is possible to effectively identify areas that are strongly affected, thereby further locating pest and disease problems. At the same time, energy calculation based on grayscale co-occurrence matrix data can evaluate the uniformity of texture in the image. A lower energy value usually indicates that the texture is more dispersed, which is related to the feeding characteristics of pests and diseases. Energy statistics of low-energy grayscale co-occurrence matrix data help to screen out areas with more dispersed textures and greater pest impact, providing important clues for pest control. By performing intersection operations on low-energy grayscale symbiosis matrix data and high-entropy grayscale symbiosis matrix data, the areas on the leaves affected by pests can be accurately identified, and the feeding mark data can be further extracted to provide a basis for the type and degree of analysis of pests. Finally, through the slender shape recognition technology, linear feeding marks can be effectively distinguished, and the specific manifestations of pests and diseases can be accurately located, providing detailed and clear data support for subsequent agricultural management decisions. The implementation of this series of steps has effectively improved the accuracy, real-time and automation level of pest and disease monitoring in agricultural production management, and provided a more scientific and accurate technical means for intelligent agricultural decision-making.

[0051] Optionally, step S23 is specifically:

[0052] Step S231: collecting images of crop physiological data to obtain crop physiological images;

[0053] Step S232: filtering and denoising the crop physiological image to obtain a crop physiological denoised image;

[0054] Step S233: performing wormhole edge detection according to the crop physiological denoising image, thereby obtaining wormhole edge data;

[0055] Step S234: performing an expansion operation on the wormhole edge data to obtain wormhole edge expansion data;

[0056] Step S235: calibrating the wormhole region of the crop physiological denoised image according to the wormhole edge expansion data, thereby obtaining wormhole region data;

[0057] Step S236: performing quantity statistics according to the wormhole area data, thereby obtaining wormhole quantity data;

[0058] Step S237: Calculate the area according to the wormhole area data, thereby obtaining the wormhole area data;

[0059] Step S238: Perform wormhole feature fusion according to the wormhole area data and the wormhole quantity data to obtain wormhole data.

[0060] The present invention collects images of crop physiological data, and the system can effectively obtain the physiological status information of crops, laying a foundation for subsequent analysis. In the collected physiological images, there are noises and unnecessary interferences, so through filtering and denoising processing, unnecessary interference information can be removed, and clearer physiological characteristics of crops can be retained, thereby ensuring the accuracy of subsequent analysis. Then, using the wormhole edge detection technology, the wormhole traces in the crop physiological image can be accurately identified, providing a basis for further analysis of pests. After the expansion operation, the edge of the wormhole is expanded, making the detected wormhole area more obvious, and providing better data support for subsequent calibration operations. According to the expanded wormhole edge data, the wormhole area in the crop physiological image can be accurately calibrated to obtain accurate wormhole area data, thereby providing a key basis for further pest analysis. By counting the number of wormhole areas, the severity of pests can be evaluated, helping farmers to monitor the spread of pests in real time and take corresponding prevention and control measures. At the same time, the area calculation of the wormhole area helps to evaluate the scale of pests and provide the degree of crop damage, thereby providing a reference for crop health management. Finally, combining the number and area data of wormholes for feature fusion can fully reflect the comprehensive characteristics of pests, provide more comprehensive pest diagnosis information, and help agricultural managers make accurate decisions in a timely manner. The implementation of this series of steps effectively improves the monitoring accuracy and real-time performance of crop health status, promotes the refinement and intelligence of agricultural management, further optimizes the utilization of resources, and reduces the waste of pesticides and fertilizers.

[0061] Optionally, step S3 specifically includes:

[0062] Step S31: acquiring farm remote sensing data, and collecting crop spectra according to the farm remote sensing data, thereby obtaining crop spectral data;

[0063] Step S32: dividing the crop spectral data into bands, thereby obtaining green light band data and short-wave infrared band data;

[0064] Step S33: performing normalized moisture index calculation according to the green light band data and the short-wave infrared band data, thereby obtaining normalized moisture index data;

[0065] Step S34: performing water stress identification according to the normalized water index data, thereby obtaining water stress data;

[0066] Step S35: obtaining crop water stress threshold data;

[0067] Step S36: dividing the water stress data according to the crop water stress threshold data, thereby obtaining high water state data and low water state data;

[0068] Step S37: maintaining the crop irrigation plan for the high moisture state data, thereby obtaining the crop high moisture state irrigation plan data;

[0069] Step S38: increasing the irrigation frequency for the low moisture state data, thereby obtaining the irrigation frequency data for the low moisture state of the crop;

[0070] Step S39: Integrate the irrigation plan according to the crop high moisture state irrigation plan data and the crop low moisture state irrigation frequency data to obtain the crop irrigation plan data, and evaluate the moisture state according to the crop irrigation plan data to obtain the crop moisture state data.

[0071] The present invention obtains farm remote sensing data and performs spectral acquisition, and the system can accurately capture the spectral information of crops, providing an important data source for subsequent water status analysis. The band division of spectral data helps to extract specific green light band and short-wave infrared band data, which play an important role in analyzing the water status and health status of crops. Calculating the normalized moisture index based on these band data can help evaluate the moisture content and moisture status of crops, and then reveal whether the crops are in a state of water stress. Through the water stress identification technology, it is possible to timely identify whether crops are under water stress, help agricultural managers identify the severity of water problems, and thus avoid crops from being affected by drought or excessive water. According to the set water stress threshold data, water stress is divided, and high water status and low water status can be accurately distinguished, providing hierarchical data support for subsequent irrigation management. In a high water state, formulating and maintaining a suitable irrigation plan can ensure that crops grow under suitable water conditions and avoid problems such as root hypoxia caused by excessive water; while in a low water state, appropriately increasing the irrigation frequency can effectively avoid crop drought, ensure sufficient water supply for crops, and promote healthy growth. Finally, by integrating the irrigation plan data of high and low moisture states, an overall crop irrigation plan is formed, and moisture state assessment is performed to ensure that agricultural management is more targeted and accurate. This series of steps effectively promotes the intelligent management of agricultural production, reduces resource waste, optimizes crop moisture management, and promotes the refinement and automation of agricultural management through real-time monitoring and dynamic adjustment of irrigation strategies.

[0072] Optionally, step S5 specifically includes:

[0073] Step S51: dividing the area according to the visualized data of the area affected by pests and diseases, so as to obtain data of areas with high incidence of pests and diseases;

[0074] Step S52: optimizing the pesticide spraying dosage for the pest-prone areas data, thereby obtaining the pesticide spraying dosage optimization data;

[0075] Step S53: performing crop density statistics on the pest-prone area data, thereby obtaining pest-prone crop density data;

[0076] Step S54: setting a density threshold according to the density data of crops with high incidence of insect pests, thereby obtaining a high incidence density threshold of insect pests;

[0077] Step S55: Density monitoring is performed on the visualized data of the pest-affected area according to the pest high-incidence density threshold, so as to obtain data of potential pest high-incidence areas and data of areas that have not reached the pest density threshold;

[0078] Step S56: increasing the frequency of pesticide spraying according to the data of potential high-incidence areas of insect pests, thereby obtaining pesticide spraying frequency data;

[0079] Step S57: reducing the amount of pesticide spraying according to the data of the area that does not reach the pest density threshold, thereby obtaining pesticide spraying amount data;

[0080] Step S58: constructing a pesticide intelligent spraying management strategy based on the pesticide spraying dosage optimization data, the pesticide spraying frequency data, and the pesticide spraying amount data, thereby obtaining pesticide spraying intelligent management data.

[0081] The present invention can effectively identify and locate areas with high incidence of pests by performing visual data analysis and regional division according to the areas affected by pests and diseases, providing an important basis for subsequent precise pesticide management. Further optimizing the pesticide spraying dosage in areas with high incidence of pests can ensure that the amount of pesticide used is sufficient to suppress pests without wasting resources, reducing the cost of pesticide use and the environmental burden. At the same time, by statistically analyzing the crop density in areas with high incidence of pests, the crop density in different areas can be clearly understood, helping to identify which areas are more susceptible to pests due to excessive density. Based on these density data, the high-incidence density threshold of pests is set to make pesticide spraying decisions more accurate and avoid unnecessary spraying or missed spraying. By setting the density threshold to further monitor the areas affected by pests and diseases, it is possible to distinguish between areas with high incidence of potential pests and areas that have not reached the pest density threshold, providing reliable data support for subsequent precise policy implementation. In areas with high incidence of potential pests, increasing the frequency of pesticide spraying can effectively suppress the spread of pests and ensure that crops are protected from pest threats, while reducing the amount of pesticide spraying in areas that have not reached the pest density threshold can help save resources and reduce environmental pollution. Ultimately, by comprehensively optimizing the dosage, frequency, and amount of pesticide spraying to build an intelligent spraying management strategy, the spraying plan can be dynamically adjusted according to the pest risk in different regions, ensuring the refinement and efficiency of agricultural management, reducing the use of pesticides, improving the health of crops, and reducing production costs at the same time. This series of steps effectively improves the intelligence and automation level of pesticide management, contributes to the rational use of resources and environmental protection, and promotes the transformation of agricultural production towards sustainable and green development.

[0082] Optionally, this specification also provides a smart agricultural information big data intelligent collection and management system, which is used to execute the smart agricultural information big data intelligent collection and management method as described above, and the smart agricultural information big data intelligent collection and management system includes:

[0083] A crop growth analysis module is used to obtain farm sensor data and perform crop growth analysis based on the farm sensor data to obtain crop growth data;

[0084] The pest and disease detection module is used to extract crop health status characteristics and crop physiological characteristics according to crop growth data, so as to obtain crop health status data and crop physiological data; perform pest and disease detection according to crop health status data and crop physiological data, so as to obtain crop pest and disease data;

[0085] A crop moisture status assessment module is used to obtain farm remote sensing data, and perform crop multispectral analysis based on the farm remote sensing data, thereby obtaining crop multispectral data, and perform crop moisture status assessment based on the crop multispectral data, thereby obtaining crop moisture status data;

[0086] The pest and disease affected area visualization module is used to perform correlation analysis on crop moisture status data based on crop pest and disease data, thereby obtaining crop pest and disease-moisture status data; and to visualize the pest and disease affected area based on the crop pest and disease-moisture status data, thereby obtaining pest and disease affected area visualization data;

[0087] The intelligent management module for pesticide spraying is used to perform intelligent management of pesticide spraying based on the visualized data of the area affected by pests and diseases, thereby obtaining intelligent management data for pesticide spraying.

[0088] The smart agricultural information big data intelligent collection and management system of the present invention can realize any smart agricultural information big data intelligent collection and management method of the present invention, and is used to combine the operation and signal transmission medium between various modules to complete the smart agricultural information big data intelligent collection and management method. The internal modules of the system cooperate with each other to optimize the pesticide spraying strategy and improve the efficiency of agricultural production management. BRIEF DESCRIPTION OF THE DRAWINGS

[0089] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments thereof made with reference to the following drawings:

[0090] Figure 1 This is a schematic diagram of the steps of the intelligent collection and management method of smart agricultural information big data of the present invention;

[0091] Figure 2 Detailed step flow diagram of step S17 in the present invention;

[0092] Figure 3 Detailed step flow diagram of step S2 in the present invention.

[0093] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0094] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0095] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0096] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0097] To achieve this, please refer to Figures 1 to 3 The present invention provides a method for intelligent collection and management of smart agricultural information big data, the method comprising the following steps:

[0098] Step S1: acquiring farm sensor data, and performing crop growth analysis based on the farm sensor data, thereby obtaining crop growth data;

[0099] In this embodiment, farm data is collected in real time through a variety of sensors installed in the farm (such as temperature and humidity sensors, light sensors, soil moisture sensors, etc.). The data collected by each sensor include basic parameters such as soil moisture (unit: %), temperature (unit: ℃), light intensity (unit: lux), and air humidity (unit: %). The data collection frequency of these sensors is set to once an hour, and the data is transmitted to the data processing platform through a wireless network. These data are preprocessed to remove outliers and noise data. For temperature and humidity data, a reasonable range is set, such as a temperature range of 10℃ to 40℃ and a humidity range of 20% to 95%. Any value outside this range will be regarded as an outlier and will be removed or supplemented. Then, based on these data, the crop growth model (such as calculating the crop growth index based on temperature and light) is used for analysis to obtain growth data such as the growth stage, health status, and expected yield of the crop. In this step, the crop growth index (such as the NDVI index) is calculated by light intensity and temperature and humidity data. The basis for dividing the crop growth stage is the cumulative temperature value (unit: day·℃), and its threshold is set according to different crop types. Ultimately, crop growth data is generated, including the crop’s current growth stage, health status assessment, expected harvest date, etc.

[0100] Step S2: extracting crop health status characteristics and crop physiological characteristics according to crop growth data, thereby obtaining crop health status data and crop physiological data; performing pest and disease detection according to the crop health status data and crop physiological data, thereby obtaining crop pest and disease data;

[0101] In this embodiment, the health status extraction is performed by collecting the appearance image of the crop, and a high-resolution camera (such as a 50 million pixel image collector) is used to shoot the leaves, branches, fruits and other parts of the crop, and the image resolution is set to 4000×3000 pixels. The captured image will be converted into a grayscale image, and the texture characteristics, color distribution and other health status indicators of the leaves will be extracted through the image processing algorithm. For the extraction of physiological characteristics of crops, the chlorophyll content information of the plant is collected, and a chlorophyll meter (such as SPAD-502) is used for measurement. The chlorophyll value is set in the range of 0 to 100, and the photosynthesis efficiency of the crop is inferred by the measured value. After the integration of health status and physiological data, pests and diseases are detected. At this time, by using the known crop pest and disease image database and image matching algorithm, by comparing the crop health status image with the pest and disease sample library, it is detected whether there are pest and disease characteristics, such as whether there are worm holes and food marks on the leaves. Image recognition algorithms (such as edge detection and region growth algorithms) are used to identify pest areas, and the severity of the pests is calculated based on the size and distribution of the pests.

[0102] Step S3: acquiring farm remote sensing data, and performing crop multispectral analysis based on the farm remote sensing data, thereby obtaining crop multispectral data, and performing crop moisture status assessment based on the crop multispectral data, thereby obtaining crop moisture status data;

[0103] In this embodiment, a multispectral sensor carried by a remote sensing satellite or an unmanned aerial vehicle (such as the OLI sensor of Landsat 8) is used to regularly monitor the farm and obtain multispectral data of crops. Each remote sensing data set includes reflectance values ​​of red light, green light, blue light, near infrared and other bands. When collecting data, the band frequency is set to once a month, and it is ensured that the data acquired in the same time period can minimize the impact of environmental factors on the data. For the collected multispectral data, the crop health is analyzed by calculating the vegetation index (such as NDVI). The normalized difference water index (NDWI) formula is used to evaluate the moisture status of the multispectral data:

[0104]

[0105] Among them, NIR is the reflectance of the near-infrared band, and SWIR is the reflectance of the short-wave infrared band. The range of NDWI values ​​is between -1 and +1. When the NDWI value is close to 1, it means that the crops are in a good water state; when the NDWI value is close to 0, it means that the crops are in a drought state. At this time, water stress identification determines whether the crops need irrigation or water supplementation by comparing with historical data and combining the water threshold (such as NDWI value below 0.1 is a drought state).

[0106] Step S4: performing correlation analysis on the crop moisture status data according to the crop pest data, thereby obtaining crop pest-moisture status data; visualizing the pest-affected area according to the crop pest-moisture status data, thereby obtaining pest-affected area visualization data;

[0107] In this embodiment, the pest data and the water state data are subjected to association analysis. The analysis is carried out by establishing a weighted association model. There is a certain causal relationship between the water state and the occurrence of pests and diseases. For each piece of farmland, the water state (such as drought, medium, and over-wet) is divided by region and the frequency of occurrence of pests and diseases is associated. The association analysis uses a weighted coefficient (for example, the weight coefficient of the water state and the pest is set to 0.7, and the weight of the impact of the pests and diseases is 0.3). The analysis result can output the relationship between the water state and the pest density in each region, and obtain the crop pest-water state data. According to the obtained crop pest-water state data, a GIS visualization technology is used to generate a pest-affected area map. At this time, by comparing the water data and the pest density data in the region, the high-incidence area of ​​pests and diseases (for example, the water state is "drought" and the pest density value is higher than a certain threshold) and the low-incidence area of ​​pests and diseases are drawn, and a visualization map of the pest-affected area is output. This map can accurately indicate which areas of crops are affected by pests, thereby providing a basis for the next step of pesticide spraying decision.

[0108] Step S5: Perform intelligent management of pesticide spraying according to the visualized data of the area affected by pests and diseases, thereby obtaining intelligent management data of pesticide spraying.

[0109] In this embodiment, the area is divided according to the visualized data of the pest-affected area. Different pesticide spraying strategies are set by separating the high-incidence areas of pests and diseases from the low-incidence areas. For areas with high pest incidence, the spraying dose is set to a higher level (for example, the spraying amount is 30 liters of pesticide per hectare), while for areas with low pest incidence, the spraying amount is set to a lower level (10 liters of pesticide per hectare). The spraying frequency is adjusted according to the real-time density of pests, and the pest density threshold (such as the density of crop plants with pest density greater than 50%) determines the frequency and dose of spraying. A pesticide spraying dose optimization algorithm is used to adjust the spraying path and dose to ensure the coverage effect of pesticides while reducing unnecessary waste. This process uses a GPS positioning system and automated spraying equipment (such as a drone sprayer) to accurately perform spraying tasks. After pesticide spraying, real-time feedback data (such as crop health data and pest density changes) is collected, and the spraying strategy is optimized and adjusted according to the feedback data to ensure intelligent management of pesticide spraying.

[0110] Optionally, step S1 specifically includes:

[0111] Step S11: acquiring farm sensor data, and performing nutrient sensor feature extraction based on the farm sensor data, thereby obtaining nutrient sensor data;

[0112] In this embodiment, it is necessary to obtain relevant nutrient data through soil sensors installed in the farm. These sensors can measure the concentration of various nutrients in the soil, such as nitrogen (N), phosphorus (P), potassium (K), etc. Nutrient sensors are usually electrochemical sensors, and their working principle is based on ion-selective electrodes, and the nutrient level is calculated by measuring the ion concentration in the soil solution. In this process, a suitable soil sensor is selected, such as a soil nitrogen, phosphorus and potassium concentration sensor, and the sampling period of each sensor is 30 minutes, and its threshold is adjusted according to the soil type (for example, the soil nitrogen concentration is between 0.1 and 0.5g / kg, and the soil potassium concentration is between 0.3 and 1.0g / kg). After data acquisition, the data is uploaded to the central database for further processing through a data transmission device (such as LoRa or NB-IoT communication module). After acquiring the data, the nutrient sensing data is filtered by a signal processing algorithm to remove the noise caused by current fluctuations. For each sensor, the data collected each time will be averaged five data points to ensure the stability of the results. Finally, the nutrient sensing data includes the concentration data of nutrients such as nitrogen, phosphorus, and potassium obtained by each sensor, as well as the corresponding timestamp information.

[0113] Step S12: Counting soil nutrient element concentrations according to the nutrient sensing data, thereby obtaining soil nutrient element concentration data;

[0114] In this embodiment, the statistics of soil nutrient element concentrations are based on the nutrient sensor data obtained in step S11. The concentration data of each nutrient is processed using a statistical analysis method. Taking nitrogen (N) as an example, nitrogen concentration data for a certain period of time (such as once a month) is collected, its change trend is analyzed, and the average value, standard deviation, maximum value and minimum value of the nitrogen concentration are calculated. For example, the standard threshold value of nitrogen concentration is set to 0.1g / kg to 0.5g / kg. If the nitrogen concentration deviates from this range, it is regarded as an indication of insufficient or excessive soil nutrients. The data processing method used includes data cleaning and outlier removal. For example, the outlier judgment threshold of the data is set to ±3 times the standard deviation, and the values ​​beyond this range are regarded as abnormal data for removal. The nutrient concentration of each soil sample area will be counted separately, and the specific concentration data of nutrients such as nitrogen, phosphorus and potassium in each piece of soil will be obtained by analyzing each area. These statistical data include the nutrient uniformity of soil in different regions, areas with uneven nutrient distribution, etc.

[0115] Step S13: Acquire crop nutrient requirement data and crop growth stage data;

[0116] In this embodiment, the nutrient requirement data of crops is obtained through agricultural literature and crop growth models, and the model specifies the ideal concentrations of nutrients such as nitrogen, phosphorus, and potassium required by different crops at different growth stages (such as germination, branching, flowering, and fruiting). For example, corn has a higher nitrogen demand in the early growth stage, while its phosphorus demand increases during the flowering period. The crop growth stage data is obtained through sensors in the farm. The growth stage of the crop is calculated using environmental factors such as temperature, humidity, and light intensity, combined with a crop growth model (such as a temperature accumulation model). For example, within the temperature range, when the cumulative value of daily temperature reaches a certain threshold (such as 300°C·day), the crop enters the flowering period; and with the cooperation of soil moisture and light intensity, it is determined whether the crop enters the harvest period. The crop growth model is used to monitor and calibrate the crops to ensure the accuracy of its growth stage data. The data collection frequency is set to once an hour. After preprocessing, the nutrient requirement data of all growth stages and the current growth stage of the crop are summarized to obtain real-time crop nutrient requirement data and growth stage data.

[0117] Step S14: constructing a crop nutrient requirement model according to the crop nutrient requirement data and the crop growth stage data, thereby obtaining a crop nutrient requirement model;

[0118] In this embodiment, the model uses multiple regression analysis or other data-driven methods to analyze the demand for different nutrients at each growth stage. For example, in the initial growth stage, crops have a greater demand for nitrogen and phosphorus, while in the flowering stage, the demand for potassium increases. By analyzing the correlation between different nutrients and crop growth stages, the nutrient demand ratio for each stage is obtained. Use data analysis tools (such as MATLAB, Python's pandas library) to model the nutrient demand for each stage. First, take the various growth stages of crops (such as germination stage, flowering stage, etc.) as independent variables, and take the required concentrations of nitrogen, phosphorus, and potassium at each stage as dependent variables, and use a regression algorithm for fitting. Set the model threshold (such as R 2 The model is valid when the value reaches 0.95 or above, and is constantly corrected according to real-time monitoring data. The resulting nutrient requirement model can give the concentration range of different nutrients required for each stage and each crop, and adjust the demand for each nutrient according to real-time data to ensure the optimal nutritional status of crop growth.

[0119] Step S15: identifying crop nutrient ratio imbalance based on the soil nutrient element concentration data according to the crop nutrient requirement model, thereby obtaining crop nutrient ratio imbalance data;

[0120] In this embodiment, the soil nutrient concentration data is compared with the crop nutrient requirement model to identify the imbalance of nutrient ratios. By calculating the deviation between the actual concentration of each nutrient (such as nitrogen, phosphorus, and potassium) and the recommended concentration in the model, the imbalance degree of each nutrient is calculated. If the concentration of a certain nutrient exceeds the recommended range of the model, the area is regarded as a nutrient imbalance area. First, set the ideal concentration range of each nutrient (such as 0.2g / kg to 0.4g / kg for nitrogen, 0.1g / kg to 0.3g / kg for phosphorus, and 0.3g / kg to 0.8g / kg for potassium), then calculate the deviation of each nutrient according to the actual measured value of the soil, and set the deviation threshold (such as when the deviation exceeds ±10%, it is determined to be a nutrient imbalance). By calculating the nutrient deviation of each area, it is found that the nutrient ratio of which areas is unbalanced, thereby obtaining the nutrient ratio imbalance data of crops.

[0121] Step S16: extracting chlorophyll meter features according to farm sensor data, thereby obtaining chlorophyll meter data;

[0122] In this embodiment, crops in the farm are regularly tested by a chlorophyll meter (such as SPAD-502) to obtain chlorophyll content data of crop leaves. The chlorophyll meter measures the light reflection characteristics of the leaves and estimates the chlorophyll content of the leaves. Each measurement result of the process is output in real time by the chlorophyll meter in SPAD values, usually ranging from 30 to 80 SPAD units. The data collection frequency is once a day, and different areas of crops are selected for testing to ensure the representativeness of the chlorophyll data. After each measurement, the chlorophyll content value is recorded and combined with the soil nutrient data of the area to obtain the chlorophyll meter data for each crop area. All data is stored in the cloud platform for subsequent analysis and adjustment of nutrient management strategies for crops.

[0123] Step S17: Analyze the chlorophyll content of crops on the chlorophyll meter data to obtain the chlorophyll content of crops;

[0124] In this embodiment, the chlorophyll meter data is analyzed, specifically by calculating the chlorophyll content of the leaves of crops in each region to determine the photosynthetic efficiency of the crops. The collected chlorophyll meter data is used to calculate the average chlorophyll content of the crops in each region. If the chlorophyll content of a certain area is lower than a certain threshold (such as lower than 40SPAD units), it means that there is a nutrient deficiency or growth problem in the area. Use a data analysis tool to process all the collected chlorophyll meter data, calculate the average value of each area, and if the chlorophyll content of a certain area is significantly lower than the standard value of the area (such as less than 40SPAD units), it is marked as a chlorophyll-deficient area, and the reason is further analyzed. In the analysis process, soil nutrient concentration data, environmental factors and other data are combined to find out the key factors affecting the chlorophyll content.

[0125] Step S18: Perform crop growth integration according to the crop chlorophyll content and crop nutrient ratio imbalance data to obtain crop growth data.

[0126] In this embodiment, the chlorophyll content of crops (step S17) and the soil nutrient imbalance data (step S15) are combined to determine whether the crops are in the best growth state. If the chlorophyll content is low and the nutrients are imbalanced, the growth environment of the crops is optimized by adding certain nutrients or adjusting the irrigation strategy. Using a comprehensive model, the chlorophyll content of each region is combined with the nutrient imbalance data, and the comprehensive growth status of the crops is obtained by numerical weighting or decision tree algorithm. Finally, the obtained crop growth data includes key information such as the health status, nutrient requirements, and photosynthesis efficiency of the crops in each region.

[0127] Optionally, step S17 is specifically:

[0128] Step S171: performing reflected light intensity statistics on the chlorophyll meter data to obtain reflected light intensity data;

[0129] In this embodiment, the reflected light intensity of the leaves of the crop is measured by a chlorophyll meter. The chlorophyll meter emits light of a certain wavelength (usually red light and near-infrared light) to the surface of the crop leaves, and measures the light intensity reflected from the leaf surface. The data of the reflected light intensity is obtained by the light sensor of the chlorophyll meter. The sensor records the reflected light intensity according to the intensity and angle of the light source during the detection process. The light source intensity of the chlorophyll meter is calibrated to ensure its stability. For each leaf measurement point, the sampling time interval is set to 5 seconds, and it is ensured that there are no external interferences (such as water droplets or stains) on the surface of the leaf before each measurement. The reflected light intensity recorded by the chlorophyll meter is usually the signal intensity in watts (W), and the measurement process stores all data according to the timestamp. The sampling range of the reflected light intensity is set to 300 to 800nm, covering the visible light and near-infrared spectrum ranges. Finally, the collected light intensity data is integrated to obtain the average reflected light intensity data of different regions.

[0130] Step S172: performing band division according to the reflected light intensity data, thereby obtaining red light band data and near infrared light band data;

[0131] In this embodiment, the reflected light intensity data obtained in step S171 is divided into bands. According to the data of the chlorophyll meter, a specific band is selected for analysis, which is usually divided into a red light band (about 620nm to 750nm) and a near infrared band (about 750nm to 900nm). This process relies on spectral analysis technology, using a bandpass filter or a spectrometer to separate the red light and near infrared light bands. The collected light intensity data is compared with the known ranges of the red light and near infrared light bands to ensure that the data can be correctly divided. For example, the red light band is set to 620nm to 750nm, and the near infrared light band is set to 750nm to 900nm. Using a bandpass filter, the measured light intensity signal is divided by band. In this process, the measurement accuracy is guaranteed, the band overlap or spectral data loss is avoided, and the divided band data is ensured to be consistent with the actual measurement value. The light intensity data of each band will be extracted and stored independently, and finally the red light band data and the near infrared light band data will be formed.

[0132] Step S173: Calculating the reflectivity of the red light band data, thereby obtaining the red light band reflectivity data;

[0133] In this embodiment, the intensity of the reflected light is measured by a light sensor in the red light band. The reflectivity is the ratio of the intensity of the reflected light to the intensity of the incident light. In order to calculate the reflectivity, the intensity of the incident light source needs to be calibrated first, which is usually set to 40 μW / cm 2 Then, the reflected light intensity at each measurement point is measured, and the reflectance is calculated based on the known incident light intensity. The reflectance value represents the ratio of the reflected light intensity on the blade surface to the incident light intensity. After each measurement, the reflectance data is recorded and statistically processed to obtain the reflectance value of each measurement point, which is finally stored in percentage form.

[0134] Step S174: calculating the reflectivity of the near-infrared light band data, thereby obtaining the near-infrared light band reflectivity data;

[0135] In this embodiment, it is necessary to ensure that the incident light intensity is within 50 μW / cm 2 Next, the reflected light intensity of each measuring point is measured by a near-infrared light sensor, and the reflectance is calculated based on the known incident light intensity. The near-infrared light reflectance value of each measuring point will reflect the reflection characteristics of the leaf surface to near-infrared light. During the measurement process, ensure that the light source intensity and sensor working state are stable to avoid external interference. All reflectance data will be recorded in real time and cleaned to ensure its accuracy. Ultimately, these reflectance data will be used for further chlorophyll content assessment.

[0136] Step S175: evaluating the chlorophyll content of crops according to the red light band reflectance data and the near infrared light band reflectance data, thereby obtaining the chlorophyll content data of crops.

[0137] In this embodiment, the reflectance data of the red light band and the near-infrared light band are used to evaluate the chlorophyll content of crops. By calculating the vegetation index and combining the reflectance of red light and near-infrared light, the chlorophyll content of the leaves can be inferred. For each measuring point, the reflectance of the red light band is compared with the reflectance of the near-infrared light band to obtain a numerical value reflecting the health status of the plant. Generally, a higher reflectance difference (for example, an NDVI value greater than 0.6) indicates a higher chlorophyll content, while a lower reflectance difference means a lower chlorophyll content. These calculation results will be used to further analyze the growth status and health level of crops, helping farmland managers adjust the nutrient requirements and fertilization strategies of crops.

[0138] Optionally, step S2 specifically includes:

[0139] Step S21: extracting crop health status characteristics and crop physiological characteristics according to crop growth data, thereby obtaining crop health status data and crop physiological data;

[0140] In this embodiment, the growth data of crops are collected, and these data can be obtained through devices such as temperature and humidity sensors, soil moisture sensors, and leaf temperature sensors. The specific content of the growth data includes parameters such as crop growth cycle, environmental factors, and climate change. Then, these data are used to extract the health status characteristics of crops, such as leaf color changes, crop height changes, and leaf number changes. The specific method is to perform time series analysis on sensor data to extract health-related parameters such as crop growth rate and leaf area index (LAI). Next, physiological characteristics are extracted, and physiological data such as crop photosynthesis efficiency, stomatal conductance, and transpiration rate are used for analysis to extract physiological characteristics closely related to health status. Finally, based on the extracted health status and physiological data, crop health status data and crop physiological data are generated for subsequent pest and disease identification and health management.

[0141] Step S22: performing pest feeding trace analysis on the crop health status data to obtain pest feeding trace data;

[0142] In this embodiment, the image or sensor data about the leaf damage in the crop health status data is used to perform pest feeding trace analysis. The image data can be obtained by a high-definition camera or a near-infrared optical sensor, and the sensor can capture the subtle damage information on the leaf surface. Through the image processing algorithm, the background removal, image enhancement and other pre-processing are first performed. Then, the edge detection algorithm (such as Canny edge detection) is used to extract the outline of the pest feeding trace on the leaf surface, and the regional growth algorithm or connected region analysis is further used to identify the feeding trace area, and the number, size, distribution and other characteristics of the feeding trace are calculated. The parameters of each feeding trace will be combined with the growth state of the crop, and by setting a threshold, a significant pest feeding trace area is identified. Finally, through statistical analysis, the pest feeding trace data is obtained for subsequent pest and disease analysis.

[0143] Step S23: performing wormhole analysis on crop physiological data to obtain wormhole data;

[0144] In this embodiment, wormhole analysis is performed based on the physiological data of crops. Physiological data include information such as the growth status of crops, stomatal conductance, and leaf morphology. First, use a high-resolution imaging device (such as an infrared thermal imager or a microscope) to scan the leaves of crops to obtain structural data on the surface and inside of the leaves. Then, image processing algorithms, such as morphological operations and threshold segmentation, are used to extract wormhole areas on the leaf surface. These wormholes are caused by insect pests, climate change or diseases. During the analysis, the size, shape, depth and other characteristics of the wormholes are combined with the growth data of the crops for feature extraction. By comparing the wormhole data of different crops and different growth stages, potential wormhole formation areas are identified, and the wormhole data is output for subsequent pest and disease intervention measures.

[0145] Step S24: Integrate the characteristics of crop pests and diseases based on the pest feeding trace data and the wormhole data, so as to obtain crop pest and disease data.

[0146] In this embodiment, the insect pest feeding trace data and the wormhole data are integrated. During the integration process, the feeding trace data and the wormhole data are first matched by matching the timestamp or spatial position. Then, the two types of data are associated and analyzed based on statistical methods (such as correlation analysis or regression analysis) to identify patterns and trends related to pests and diseases. For example, the law of pest occurrence can be extracted by comparing the feeding traces and wormholes of pests and diseases in different crops or growth stages. Then, thresholds are set according to these laws and each data point is classified. Using threshold screening, the areas and degrees of pests and diseases in crops are identified. Finally, the feeding trace data, wormhole data and other crop health-related data (such as chlorophyll content, stomatal conductance, etc.) are combined to generate complete crop pest and disease data. These data will provide a basis for subsequent crop protection measures.

[0147] Optionally, step S22 is specifically:

[0148] Step S221: collecting images of crop health status data to obtain crop health status images;

[0149] In this embodiment, the image acquisition device is used to obtain images of the health status of crops. Use a high-resolution digital camera or a near-infrared imaging device, and set appropriate shooting parameters, such as focal length, exposure time, and white balance, to ensure stable image quality. Select an appropriate acquisition angle (such as a vertical or oblique angle) and shooting distance to ensure that the image can fully cover the health status of the target area. During the image acquisition process, be sure to keep the lighting conditions stable to avoid interference from shadows or reflections. Specific lighting standards can be set, such as image acquisition under a constant light of 1000 lux. During the acquisition process, the resolution of the shooting device is adjusted according to the type, size, and growth stage of the crops. A resolution of 5000×3000 pixels is usually selected to ensure that details are clearly discernible. Ultimately, the acquired image files will be stored in a standard format (such as JPEG, TIFF, etc.) and subsequently processed.

[0150] Step S222: performing grayscale conversion according to the crop health status image, thereby obtaining a crop health status grayscale image;

[0151] In this embodiment, a grayscale conversion process is performed on the color crop health status image obtained in step S221. The grayscale value is obtained by weighted averaging the red, green and blue (RGB) channel values ​​of each pixel. The specific implementation method is to use the weighted average formula: grayscale value = 0.2989*red channel value + 0.5870*green channel value + 0.1140*blue channel value. This formula is used to convert each pixel of the image to generate a grayscale image. During the conversion process, if the color space used by the image is RGB, each pixel value after conversion will be in the range of 0 to 255, representing different grayscale levels. This process is implemented by an image processing library (such as OpenCV) in a programming language (such as Python, C++). The generated grayscale image will be stored in a grayscale image format and can be used for subsequent analysis.

[0152] Step S223: extracting leaf texture features according to the grayscale image of the crop health status, thereby obtaining leaf texture data;

[0153] In this embodiment, the grayscale image is subjected to image enhancement processing, such as histogram equalization, to enhance the contrast of texture features. Then, methods such as local binary pattern (LBP) or grayscale co-occurrence matrix (GLCM) are used to extract texture features in the image. The spatial relationship of the grayscale values ​​of pixels in the image is calculated by the GLCM method, thereby extracting statistical features that can characterize the leaf texture, such as contrast, correlation, uniformity, and entropy. During the implementation process, a set window size (such as 3×3 or 5×5) is used to slide on the image to calculate the texture features of each window area. For the LBP method, a binary pattern is generated by comparing the grayscale values ​​of pixels in the local neighborhood, and texture information is further extracted. Finally, leaf texture data is generated by the extracted texture features and stored as a feature matrix or data set for subsequent analysis.

[0154] Step S224: performing roughness statistics on the blade texture data to obtain high-roughness blade texture data;

[0155] In this embodiment, based on the extracted texture data, a roughness calculation formula is used to evaluate the roughness of the blade surface. The specific method is to use the standard deviation, volatility or other roughness indicators in the texture features to calculate the roughness of each area. Set a roughness threshold. For example, when the roughness is greater than a specific value (such as 0.2), the area is considered to be a high roughness area. Through the spatial information of the image, the roughness value is mapped to different areas of the blade to generate an image of the high roughness area. Finally, through statistical analysis, high roughness blade texture data is obtained, indicating the area on the blade surface that is damaged or affected by pests and diseases. These data will be used for subsequent pest and disease analysis.

[0156] Step S225: dividing the crop health grayscale image into regions according to the high-roughness leaf texture data, thereby obtaining high-roughness leaf texture region data;

[0157] In this embodiment, according to the set roughness threshold, the roughness value of each pixel in the image is compared with the threshold, and the pixel points belonging to the high roughness area are identified. These high roughness areas usually correspond to areas where leaves are damaged or affected by pests and diseases. Then, an image segmentation algorithm (such as threshold-based segmentation or K-means clustering) is used to distinguish the high roughness areas from other areas in the grayscale image. The high roughness areas are automatically marked by the algorithm and compared with the original image. Finally, high roughness leaf texture area data is generated, and the spatial position, area, shape and other information of these areas are recorded. This data is used for subsequent pest detection and health assessment.

[0158] Step S226: performing linear food mark recognition according to the leaf texture data, thereby obtaining linear food mark data;

[0159] In this embodiment, the slender structures on the leaves, i.e., linear food scars, are identified by image filtering and edge detection algorithms (such as Canny edge detection). Then, morphological operations (such as erosion and dilation) are used to process the image to enhance the significance of the linear food scars. In order to accurately identify linear food scars, it is necessary to set the geometric characteristic parameters of the food scars. For example, the width of a linear food scar is usually between 0.1 and 2 mm, and a length exceeding a certain threshold (such as 10 mm) can be identified as a valid food scar. Using these criteria, all qualified linear food scar areas are screened out. Ultimately, the identified linear food scar data will include information such as the number, length, and location of the food scars, and will be stored as a data set for subsequent analysis.

[0160] Step S227: performing an insect feeding trace intersection operation based on the linear feeding trace data and the high-roughness leaf texture area data to obtain the insect feeding trace data.

[0161] In this embodiment, the linear food mark data is spatially compared with the high-roughness leaf texture area data to check whether each linear food mark is located in the high-roughness area. A spatial overlap algorithm is used to calculate the intersection of the food mark and the rough area. When the linear food mark is located in the high-roughness area, the area is considered to be affected by insect pests. An intersection area threshold is set. When the intersection area exceeds a certain proportion (for example, 50%), the area is considered to be an insect pest food mark area. Through the intersection operation, the insect pest food mark data is finally obtained, the location, quantity, area and other information of the insect pest food marks are recorded, and a data report is generated. These data provide an important basis for the subsequent management of crop diseases and pests.

[0162] Optionally, step S226 is specifically:

[0163] Gray-level co-occurrence matrix calculation is performed according to leaf texture data, thereby obtaining gray-level co-occurrence matrix data;

[0164] In this embodiment, the grayscale co-occurrence matrix (GLCM) is calculated for the acquired leaf texture data. The leaf texture data is converted into a grayscale image, and then a sliding window of a fixed size (for example, 5×5 pixels or 7×7 pixels) is selected in the image, and the co-occurrence frequency of each pair of grayscale values ​​is calculated within the window. The calculation method of the co-occurrence matrix is ​​based on the relative spatial position relationship between the grayscale value pairs. Common calculation parameters include offset (such as distance 1, angle 0°) and the size of the distance window. The calculated grayscale co-occurrence matrix describes the spatial arrangement information of different grayscale pairs in the image, forming a symmetric matrix. Each element represents the frequency of a specific grayscale value pair in the image. The selected window size and offset both affect the calculation results of the co-occurrence matrix. Usually, a distance of 1 pixel and angles of 0°, 45°, 90°, and 135° are selected for multi-angle calculations. Through these grayscale co-occurrence matrix data, the spatial arrangement characteristics of the texture can be effectively described.

[0165] Perform entropy statistics on the gray-level co-occurrence matrix data to obtain high-entropy gray-level co-occurrence matrix data;

[0166] In this embodiment, entropy is an indicator to measure the complexity of image texture. The higher the entropy value, the more complex the texture structure in the image. In actual operation, the grayscale co-occurrence matrix is ​​first normalized so that the sum of all elements in the matrix is ​​1. Next, the probability of each element of the matrix is ​​calculated, and the entropy value is calculated based on the probability value. The calculation of the entropy value reflects the spatial distribution characteristics between gray levels. A higher entropy value usually means that there are more texture details or changes in the image. In order to extract high entropy areas, an entropy threshold value (for example, 7.0) is set. When the entropy value is higher than this threshold value, it is considered that the area has a higher texture complexity or pest and disease characteristics. Finally, through the statistical analysis of the entropy value, those areas with more complex textures and affected by pests and diseases are screened out, and the corresponding high entropy grayscale co-occurrence matrix data are generated. These data will help further analyze the texture features in the image, especially for the detection and evaluation of pests and diseases.

[0167] Energy calculation is performed according to the gray-level co-occurrence matrix data, thereby obtaining gray-level co-occurrence matrix energy data;

[0168] In this embodiment, energy is a measure of texture consistency and indicates the uniformity of texture. In order to perform energy calculation, the grayscale co-occurrence matrix is ​​first normalized. After normalization, the value of each element represents the probability of occurrence of the corresponding grayscale pair. The energy value of the image is obtained by calculating the square of each grayscale co-occurrence matrix element and summing them. The energy value reflects the consistency of the texture. Areas with higher energy values ​​mean that the texture of the image is more uniform, while lower energy values ​​indicate that the texture changes are larger or more complex. During implementation, high-energy areas can be screened by setting an energy threshold (such as 0.5). These areas usually show more uniform textures and are associated with healthy crops or areas without pests and diseases. In this way, the calculated grayscale co-occurrence matrix energy data can help further analyze the texture features of the image and identify healthy areas or damaged areas.

[0169] Perform energy statistics on the gray-level co-occurrence matrix energy data to obtain low-energy gray-level co-occurrence matrix data;

[0170] In this embodiment, the energy value of each pixel in the image is counted, and the set energy threshold (such as 0.3 or 0.4) is used to screen out areas with energy lower than this value and mark them as low-energy areas. The selection of this energy threshold is based on the texture characteristics of the actual image, and can distinguish areas with more complex textures or more serious pests and diseases. In specific operations, the energy data is first globally counted to calculate the energy distribution of the overall image, and then the areas below the set threshold are marked. Finally, low-energy grayscale co-occurrence matrix data is generated, indicating areas in the image with rough, uneven textures or affected by pests and diseases.

[0171] The food trace data is obtained by performing intersection operation on the low-energy gray-level co-occurrence matrix data and the high-entropy gray-level co-occurrence matrix data;

[0172] In this embodiment, the low-energy grayscale co-occurrence matrix data and the high-entropy grayscale co-occurrence matrix data are spatially overlapped at the pixel level. The labels of each pixel in the two types of data are compared. If a pixel is marked as a pest area in both types of data, the area is considered to be a food trace area. The criterion for the intersection operation is that the overlapping area of ​​the pixel in the low-energy and high-entropy regions is greater than a set threshold (for example, 10% or 20%), that is, the area is considered to be a food trace area. The intersection operation is implemented using an "AND" operation through image processing tools (such as OpenCV). Finally, image data representing the food traces of pests and diseases is obtained, and the position, shape, and size of the food trace area are marked and stored in the form of data for subsequent analysis.

[0173] The food trace data is subjected to elongated shape recognition to obtain linear food trace data.

[0174] In this embodiment, the shape of the food marks is analyzed by morphological analysis methods (such as skeleton extraction, contour detection, etc.) to identify slender linear features. In specific operations, the geometric standards of linear food marks are set. Usually, the width of the linear food marks is between 0.1 and 1 mm, and the length is greater than the set threshold (such as 10 mm) and can be regarded as linear food marks. By calculating the aspect ratio of each food mark area, the food mark area with an aspect ratio greater than a certain value (such as 2:1) is identified. Through the slender shape recognition algorithm, the area that meets the linear food mark standard is screened out and marked as linear food mark data. Finally, image data or data sets containing linear food marks are generated, recording detailed information such as the length, width, and position of the food marks.

[0175] Optionally, step S23 is specifically:

[0176] Step S231: collecting images of crop physiological data to obtain crop physiological images;

[0177] In this embodiment, crop physiological images are obtained by performing image acquisition on crop physiological data. A high-resolution digital camera or professional imaging equipment is used in the image acquisition process, and the resolution of the equipment must ensure that the physiological characteristics of crops can be clearly captured, such as subtle changes in leaves, the effects of pests and diseases, etc. The focal length and aperture of the camera are set to ensure that the crops within the shooting range are accurately captured. During the acquisition process, the image acquisition angle and lighting conditions must be stable to eliminate errors caused by changes in shooting angle or light. The acquired images need to be saved in a standard image format (such as PNG or TIFF), and basic information such as the time and place of acquisition are marked for subsequent analysis and comparison. In this process, ensure that the storage and processing of all image data do not lose any details, and perform data backup as needed.

[0178] Step S232: filtering and denoising the crop physiological image to obtain a crop physiological denoised image;

[0179] In this embodiment, the physiological denoised image of crops is obtained by filtering and denoising the physiological image of crops. The denoising operation adopts the median filtering technology, which can effectively eliminate the salt and pepper noise in the image. First, select a suitable window size, such as a 3x3 or 5x5 window, analyze each pixel in the image, and replace the current pixel value with the median of all pixel values ​​in the window. This operation can smooth the image and remove local noise. During the image denoising process, in order to ensure that the image details are not lost, it is necessary to adjust the size of the filter window and the number of filters to avoid excessive smoothing. In actual operation, it is necessary to monitor the quality of the processed image and visually evaluate the denoising effect to ensure that the denoising process can both remove noise and retain the physiological characteristics of crops.

[0180] Step S233: performing wormhole edge detection according to the crop physiological denoising image, thereby obtaining wormhole edge data;

[0181] In this embodiment, wormhole edge detection is performed based on the physiological denoising image of crops to obtain wormhole edge data. The Canny edge detection algorithm is used for edge detection. First, the denoised image is Gaussian blurred, and a Gaussian filter with a standard deviation of 1.0 is used to smooth the image to reduce the impact of noise. Then the gradient of the image is calculated, and the edge is extracted according to the set low and high thresholds. Specifically, the low threshold is set to 50 and the high threshold is set to 150. Only when the gradient value is greater than the high threshold, the edge will be marked as significant. In this process, the threshold needs to be adjusted repeatedly to ensure that the detected edge details match the morphological characteristics of the wormhole. The data after edge detection will be presented in the form of a binary image, in which the edge area of ​​the wormhole is white and the rest of the area is black.

[0182] Step S234: performing an expansion operation on the wormhole edge data to obtain wormhole edge expansion data;

[0183] In this embodiment, the expansion operation is performed by using a structural element to expand the image. A 3x3 matrix is ​​selected as the structural element. In the edge image, the structural element is convolved with the image and expanded pixel by pixel. The expansion operation expands the edge area of ​​the wormhole, making the edge more obvious. The number of iterations needs to be set during the expansion process, generally 1 to 2 times. Excessive expansion will cause excessive expansion of the wormhole edge, thereby affecting subsequent calibration and analysis. Through the expanded image, the boundary of the wormhole can be seen more clearly, which is convenient for subsequent regional calibration.

[0184] Step S235: calibrating the wormhole region of the crop physiological denoised image according to the wormhole edge expansion data, thereby obtaining wormhole region data;

[0185] In this embodiment, the calibration operation is performed by superimposing the expanded wormhole edge data with the original image, and using the grayscale information in the image to determine the wormhole area. Through an image segmentation algorithm (such as a region growing algorithm), the area related to the wormhole is automatically extracted based on the expanded edge information. During the calibration process, it is necessary to set the seed point for regional growth, and the seed point can be selected from the center position of the wormhole or other obvious feature points. The calibration operation marks the wormhole area as a specific area, and finally outputs a binary image in which the wormhole area is white and other areas are black. This calibration result will serve as the basic data for subsequent wormhole quantity statistics and area calculations.

[0186] Step S236: performing quantity statistics according to the wormhole area data, thereby obtaining wormhole quantity data;

[0187] In this embodiment, the number statistics is achieved by connecting domain analysis of the wormhole area data. A connected domain calibration algorithm is used to identify each independent wormhole area in the image. Each wormhole area will be assigned a unique label, and the number of all labels can be counted to obtain the number of wormholes. In the implementation process, it is necessary to set the judgment criteria of the connected domain, and the connected area is usually determined based on the pixel adjacency relationship, such as a four-neighborhood or an eight-neighborhood. In the statistical process, ensure that all wormhole areas are correctly counted and avoid misjudgment due to noise or unclear boundaries.

[0188] Step S237: Calculate the area according to the wormhole area data, thereby obtaining the wormhole area data;

[0189] In this embodiment, the area calculation is implemented by calculating the number of pixels in the calibrated wormhole area. The area of ​​each wormhole area is equal to the number of pixels in the area. First, based on the wormhole area data calibrated in step S235, the number of pixels in each wormhole area is calculated. If high-precision area calculation is required, it can be converted according to the actual area (such as millimeters or centimeters) of each pixel to obtain the real wormhole area data. In the implementation process, it is necessary to ensure the integrity of the area during the calculation process and exclude stray pixels caused by image processing errors.

[0190] Step S238: Perform wormhole feature fusion according to the wormhole area data and the wormhole quantity data to obtain wormhole data.

[0191] In this embodiment, feature fusion is achieved by weighted averaging the number and area of ​​wormholes or other fusion methods. The weight ratio of the area and number of each wormhole can be set. If the wormhole has a larger area, it will contribute more to the final feature. The fusion process statistically processes the area and number data of the wormholes to obtain a comprehensive wormhole feature data set. This data set can be used for subsequent crop health assessments to help identify severely damaged areas. In the implementation process, feature fusion needs to set weights and calculation methods according to specific research objectives to ensure that the fused data is representative and practical.

[0192] Optionally, step S3 specifically includes:

[0193] Step S31: acquiring farm remote sensing data, and collecting crop spectra according to the farm remote sensing data, thereby obtaining crop spectral data;

[0194] In this embodiment, it is necessary to obtain remote sensing data of the farm. Remote sensing data can be collected through satellite remote sensing, unmanned aerial vehicle equipped with spectral instruments or ground sensors. When collecting, use multispectral or hyperspectral sensors to ensure that various spectral information of crops can be captured. Spectral data needs to be collected in different bands, including visible light, near infrared, short-wave infrared and other bands, especially focusing on the reflectivity changes in green light and short-wave infrared bands. The band setting of the sensor should be optimized according to the characteristics of the crops and environmental conditions to ensure sufficient spectral resolution. During the collection process, ensure the calibration of the sensor to avoid errors, and perform data storage management to ensure the integrity and accuracy of the spectral data. The collected crop spectral data will be used for subsequent band division and moisture index calculation.

[0195] Step S32: dividing the crop spectral data into bands, thereby obtaining green light band data and short-wave infrared band data;

[0196] In this embodiment, the corresponding band range is selected according to the collected crop spectral data. The green light band is generally located between 500 and 570nm in the visible spectrum range, while the short-wave infrared band is located between 1200 and 2500nm. In order to accurately divide the bands, spectral data processing software (such as ENVI, MATLAB, etc.) is used for band selection and data extraction. For the green light band, the reflectivity data within the corresponding wavelength range is extracted; for the short-wave infrared band, the reflectivity information of the corresponding band is extracted. When extracting data, the accuracy of the band division is ensured, and errors caused by sensor noise or environmental factors are eliminated. Finally, green light band data and short-wave infrared band data are generated to provide basic data for subsequent normalized moisture index calculations.

[0197] Step S33: performing normalized moisture index calculation according to the green light band data and the short-wave infrared band data, thereby obtaining normalized moisture index data;

[0198] In this embodiment, the reflectance of the green light and short-wave infrared bands is normalized by a calculation formula. The normalized moisture index formula is:

[0199]

[0200] Extract the reflectance values ​​of the green band and the short-wave infrared band and substitute them into the formula for calculation. Ensure that the band reflectance values ​​of all data are within the range of 0 to 1 to avoid deviations in the calculation results due to numerical anomalies. The calculated NDWI value represents the moisture status of the crops. The higher the NDWI value, the higher the moisture content of the crops. During the calculation process, ensure that each pixel is processed and that the output NDWI image can accurately reflect the distribution of crop moisture.

[0201] Step S34: performing water stress identification according to the normalized water index data, thereby obtaining water stress data;

[0202] In this embodiment, different levels of water stress are divided according to thresholds provided by experience or literature. For example, if the NDWI value is lower than 0.1, it means that the water stress is more serious. Conversely, when the NDWI value is higher, the crops are in a normal water state. In order to accurately identify water stress, the NDWI data needs to be processed at the pixel level, and the set threshold (such as NDWI ≤ 0.2 for high water stress, NDWI> 0.2 for low water stress) is used to judge the water stress status of each area. All pixels below the threshold are marked as "high water stress" areas, and areas above the threshold are marked as "low water stress" areas. After the processing is completed, water stress data is generated to indicate the water stress status of different regions. This data will be used for subsequent water status management and adjustment of irrigation plans.

[0203] Step S35: obtaining crop water stress threshold data;

[0204] In this embodiment, the threshold data of crop water stress is obtained. According to the type of crop, climatic conditions and growth stage, a reasonable water stress threshold is set. The setting of the threshold is usually based on long-term field experimental data or historical meteorological data, and the threshold value can be obtained through expert experience or scientific literature. For example, some crops begin to experience water stress when the NDWI value is below 0.2, or enter a state of severe water shortage when the NDWI value is below 0.15. The threshold data can be calibrated by comparing data analysis with field observations to ensure its representativeness and practicality. During the implementation process, the threshold is set according to the specific crop and soil type to ensure that the selected threshold can effectively reflect the water demand and stress status of the crop.

[0205] Step S36: dividing the water stress data according to the crop water stress threshold data, thereby obtaining high water state data and low water state data;

[0206] In this embodiment, the water stress data is divided into two categories: high water state and low water state by setting a threshold value (for example, NDWI value 0.2). In specific implementation, the NDWI value of each pixel is compared with the threshold value. If the NDWI value is higher than the threshold value, the area is marked as high water state; if the NDWI value is lower than the threshold value, it is marked as low water state. Through this division, two data sets are obtained, which respectively represent the areas where crops are in high water state and low water state. These data will provide a basis for subsequent adjustments to the irrigation plan.

[0207] Step S37: maintaining the crop irrigation plan for the high moisture state data, thereby obtaining the crop high moisture state irrigation plan data;

[0208] In this embodiment, for areas with high moisture content, the main purpose of the irrigation plan is to maintain the appropriate moisture content of the soil and avoid over-irrigation. The irrigation cycle and irrigation amount of the high moisture content area are determined according to the settings of the irrigation system and the moisture content of the area. Generally, the irrigation cycle can be set to once every 7 days, and the irrigation amount should be determined based on the soil moisture retention capacity to ensure that water resources are avoided while maintaining the appropriate humidity. During the implementation process, the working time and water flow of the water pump and irrigation equipment can be accurately adjusted through the automated control of the intelligent irrigation system to ensure moisture balance.

[0209] Step S38: increasing the irrigation frequency for the low moisture state data, thereby obtaining the irrigation frequency data for the low moisture state of the crop;

[0210] In this embodiment, for areas with low moisture, the irrigation plan needs to increase the frequency of irrigation to restore the moisture balance as soon as possible. According to the NDWI value of the low moisture area and the moisture demand of the crop, a higher frequency of irrigation is set. For example, the low moisture area needs to be irrigated every 3 days, and the irrigation amount needs to be adjusted according to the soil type and crop demand. During the implementation process, combined with the real-time data monitored by the sensor, it is ensured that the low moisture area will not be over-irrigated during the recovery process, and the irrigation amount and frequency are adjusted in real time through the intelligent control system.

[0211] Step S39: Integrate the irrigation plan according to the crop high moisture state irrigation plan data and the crop low moisture state irrigation frequency data to obtain the crop irrigation plan data, and evaluate the moisture state according to the crop irrigation plan data to obtain the crop moisture state data.

[0212] In this embodiment, the irrigation frequency and irrigation amount are integrated into a unified plan through the intelligent irrigation system, and dynamically adjusted according to the moisture status of the crops. The irrigation plan data will be updated according to the real-time moisture monitoring data to ensure that each area grows under suitable moisture conditions. Based on these irrigation plan data, the moisture status of the crops is evaluated, the irrigation strategy is adjusted, and finally the moisture status data of the crops is obtained for decision support.

[0213] Optionally, step S5 specifically includes:

[0214] Step S51: dividing the area according to the visualized data of the area affected by pests and diseases, so as to obtain data of areas with high incidence of pests and diseases;

[0215] In this embodiment, the visualization data of the pest-affected areas in the farmland are obtained through the pest monitoring system, and these data come from ground sensors, drones or satellite remote sensing images. The visualization data is usually a heat map or classification map generated by image processing software (such as ArcGIS or QGIS), which identifies the degree of impact of different pests and diseases. Different areas are divided according to the preset pest threshold. The threshold can be set based on factors such as historical pest occurrence data, crop growth stage, and climatic conditions. For example, when the pest impact intensity exceeds a certain value (such as more than 80% of the crops are damaged), the area is divided into a high-incidence area of ​​pests. Then, based on this threshold, the coordinate information of the high-incidence area of ​​pests is extracted from the visualization data and calibrated to obtain the high-incidence area data of pests.

[0216] Step S52: optimizing the pesticide spraying dosage for the pest-prone areas data, thereby obtaining the pesticide spraying dosage optimization data;

[0217] In the present embodiment, according to the area of ​​the high-incidence area of ​​insect pests and the growth of crops, in combination with the type and concentration of pests and diseases, the appropriate pesticide spraying dosage is determined. For each high-incidence area of ​​insect pests, the required pesticide spraying amount is calculated. This amount can be estimated by the standard spraying dosage formulated according to factors such as crop type, insect pest density and climatic conditions. For example, for the insect pest density of crops per square meter exceeding a certain standard (such as the number of insect pest individuals reaching 50 per square meter), it is necessary to spray a specified dose of pesticide (such as spraying 100 ml of pesticide per square meter). In the optimization process of spraying dosage, considering reducing pesticide waste and ensuring effective prevention and control, the precise control algorithm in the intelligent irrigation system or spray device is used to adjust the spraying amount. Finally, the optimized spraying dosage data is generated.

[0218] Step S53: performing crop density statistics on the pest-prone area data, thereby obtaining pest-prone crop density data;

[0219] In this embodiment, remote sensing images or high-resolution image data taken by drones are used to identify and locate crops in the farmland. The crop planting density in each high-incidence area of ​​pests is calculated by image processing algorithms (such as image segmentation, edge detection, etc.). When counting density, the following factors can be considered: crop type, row spacing, plant spacing, etc., and statistics are performed according to standard plant density units, such as the number of plants per square meter. By analyzing the crop density in each high-incidence area of ​​pests, crop density data is obtained. This data will provide a reference for subsequent optimization of pest control and pesticide spraying, especially in areas with too high density, which will require more intensive prevention and control measures.

[0220] Step S54: setting a density threshold according to the density data of crops with high incidence of insect pests, thereby obtaining a high incidence density threshold of insect pests;

[0221] In this embodiment, based on the crop density data, a suitable threshold is set to divide the density area with high incidence of insect pests. The basis for setting the threshold usually includes the growth characteristics of the crop, the normal planting density range, and the density level of historical insect pests. For example, for a certain crop, if the density exceeds 30 plants per square meter, the probability of insect pests increases significantly. By analyzing the historical data and combining the different crop varieties, a density threshold is determined (such as more than 30 plants per square meter is a high-density area). After the threshold is set, the density data is compared with the threshold to determine which areas belong to the high-incidence density area of ​​insect pests.

[0222] Step S55: Density monitoring is performed on the visualized data of the pest-affected area according to the pest high-incidence density threshold, so as to obtain data of potential pest high-incidence areas and data of areas that have not reached the pest density threshold;

[0223] In this embodiment, the set pest high-incidence density threshold is used to perform density assessment on the data of the pest-affected area. By comparing the crop density of each area with the set density threshold, it is determined which areas have crop density exceeding the threshold, and it is determined whether pests occur in these areas. These areas are potential pest high-incidence areas. If the crop density in a certain area does not reach the threshold, it is marked as an area that does not reach the pest density threshold. Through this process, two types of data can be obtained: one is the data of potential pest high-incidence areas, and the other is the data of areas that do not reach the pest density threshold. These data will provide a basis for the optimization of the frequency and amount of pesticide spraying in the next step.

[0224] Step S56: increasing the frequency of pesticide spraying according to the data of potential high-incidence areas of insect pests, thereby obtaining pesticide spraying frequency data;

[0225] In this embodiment, the area where the spraying frequency needs to be increased is calculated based on the area and pest density data of the potential high-incidence area of ​​pests. For these areas, the spraying frequency should be set according to the probability of occurrence of pests and the growth cycle of crops. For example, if the pest density in a certain area exceeds the set threshold, the spraying frequency is increased, such as spraying once a week, or adjusted in time according to the climate and crop type. During implementation, taking into account the diversity of the agricultural environment, an intelligent pesticide spraying system is used for automatic control to accurately adjust the spraying frequency and ensure timely pest control.

[0226] Step S57: reducing the amount of pesticide spraying according to the data of the area that does not reach the pest density threshold, thereby obtaining pesticide spraying amount data;

[0227] In this embodiment, for areas that do not reach the pest density threshold, the spraying amount should be appropriately reduced to avoid pesticide waste and environmental pollution. First, the set threshold is used to determine which areas need to reduce the amount of pesticide spraying. For these areas, the spraying amount can usually be reduced to 30% to 50% of the normal spraying amount, and the specific value can be adjusted according to the growth requirements of the crop and environmental factors. During the implementation process, the precise control system of the pesticide spraying equipment is used to automatically adjust the spraying amount according to the regional needs.

[0228] Step S58: constructing a pesticide intelligent spraying management strategy based on the pesticide spraying dosage optimization data, the pesticide spraying frequency data, and the pesticide spraying amount data, thereby obtaining pesticide spraying intelligent management data.

[0229] In this embodiment, various types of data are integrated, including spraying dosage optimization data, spraying frequency data, and spraying volume data in areas with high incidence of insect pests. Based on these data, an intelligent pesticide spraying management system is constructed. The system can obtain the insect pest conditions, weather changes, and crop growth conditions of farmland in real time, and automatically adjust the spraying plan based on this information. During the implementation process, an automated control system is used to flexibly adjust the spraying dosage, frequency, and volume according to different crops, climatic conditions, and insect pest intensity. Finally, intelligent management data for pesticide spraying is generated, which provides a decision-making basis for the precise management of pesticide spraying, ensuring effective pest control while reducing the amount of pesticide used and the impact on the environment.

[0230] Optionally, this specification also provides a smart agricultural information big data intelligent collection and management system, which is used to execute the smart agricultural information big data intelligent collection and management method as described above, and the smart agricultural information big data intelligent collection and management system includes:

[0231] A crop growth analysis module is used to obtain farm sensor data and perform crop growth analysis based on the farm sensor data to obtain crop growth data;

[0232] The pest and disease detection module is used to extract crop health status characteristics and crop physiological characteristics according to crop growth data, so as to obtain crop health status data and crop physiological data; perform pest and disease detection according to crop health status data and crop physiological data, so as to obtain crop pest and disease data;

[0233] A crop moisture status assessment module is used to obtain farm remote sensing data, and perform crop multispectral analysis based on the farm remote sensing data, thereby obtaining crop multispectral data, and perform crop moisture status assessment based on the crop multispectral data, thereby obtaining crop moisture status data;

[0234] The pest and disease affected area visualization module is used to perform correlation analysis on crop moisture status data based on crop pest and disease data, thereby obtaining crop pest and disease-moisture status data; and to visualize the pest and disease affected area based on the crop pest and disease-moisture status data, thereby obtaining pest and disease affected area visualization data;

[0235] The intelligent management module for pesticide spraying is used to perform intelligent management of pesticide spraying based on the visualized data of the area affected by pests and diseases, thereby obtaining intelligent management data for pesticide spraying.

[0236] The smart agricultural information big data intelligent collection and management system of the present invention can realize any smart agricultural information big data intelligent collection and management method of the present invention, and is used to combine the operation and signal transmission medium between various modules to complete the smart agricultural information big data intelligent collection and management method. The internal modules of the system cooperate with each other to optimize the pesticide spraying strategy and improve the efficiency of agricultural production management.

Claims

1. A method for intelligent collection and management of smart agricultural information big data, characterized in that: The following steps are involved: Step S1: acquiring farm sensor data, and performing crop growth analysis based on the farm sensor data, thereby obtaining crop growth data; Step S2: extracting crop health status characteristics and crop physiological characteristics according to crop growth data, thereby obtaining crop health status data and crop physiological data; performing pest and disease detection according to the crop health status data and crop physiological data, thereby obtaining crop pest and disease data; Step S3: acquiring farm remote sensing data, and performing crop multispectral analysis based on the farm remote sensing data, thereby obtaining crop multispectral data, and performing crop moisture status assessment based on the crop multispectral data, thereby obtaining crop moisture status data; Step S4: performing correlation analysis on the crop moisture status data according to the crop pest data, thereby obtaining crop pest-moisture status data; visualizing the pest-affected area according to the crop pest-moisture status data, thereby obtaining pest-affected area visualization data; Step S5: Perform intelligent management of pesticide spraying according to the visualized data of the area affected by pests and diseases, thereby obtaining intelligent management data of pesticide spraying.

2. The intelligent collection and management method of smart agricultural information big data according to claim 1 is characterized in that: Step S1 is specifically as follows: Step S11: acquiring farm sensor data, and performing nutrient sensor feature extraction based on the farm sensor data, thereby obtaining nutrient sensor data; Step S12: Counting soil nutrient element concentrations according to the nutrient sensing data, thereby obtaining soil nutrient element concentration data; Step S13: Acquire crop nutrient requirement data and crop growth stage data; Step S14: constructing a crop nutrient requirement model according to the crop nutrient requirement data and the crop growth stage data, thereby obtaining a crop nutrient requirement model; Step S15: identifying crop nutrient ratio imbalance based on the soil nutrient element concentration data according to the crop nutrient requirement model, thereby obtaining crop nutrient ratio imbalance data; Step S16: extracting chlorophyll meter features according to farm sensor data, thereby obtaining chlorophyll meter data; Step S17: Analyzing the chlorophyll content of crops on the chlorophyll meter data, thereby obtaining the chlorophyll content of crops; Step S18: Perform crop growth integration according to the crop chlorophyll content and crop nutrient ratio imbalance data to obtain crop growth data.

3. The intelligent collection and management method of smart agricultural information big data according to claim 2 is characterized in that: Step S17 is specifically as follows: Step S171: performing reflected light intensity statistics on the chlorophyll meter data to obtain reflected light intensity data; Step S172: performing band division according to the reflected light intensity data, thereby obtaining red light band data and near infrared light band data; Step S173: Calculating the reflectivity of the red light band data, thereby obtaining the red light band reflectivity data; Step S174: calculating the reflectivity of the near-infrared light band data, thereby obtaining the near-infrared light band reflectivity data; Step S175: evaluating the chlorophyll content of crops according to the red light band reflectance data and the near infrared light band reflectance data, thereby obtaining the chlorophyll content data of crops.

4. The intelligent collection and management method of smart agricultural information big data according to claim 1 is characterized in that: Step S2 is specifically as follows: Step S21: extracting crop health status characteristics and crop physiological characteristics according to crop growth data, thereby obtaining crop health status data and crop physiological data; Step S22: performing pest feeding trace analysis on the crop health status data to obtain pest feeding trace data; Step S23: performing wormhole analysis on crop physiological data to obtain wormhole data; Step S24: Integrate the characteristics of crop pests and diseases based on the pest feeding trace data and the wormhole data, so as to obtain crop pest and disease data.

5. The intelligent collection and management method of smart agricultural information big data according to claim 4 is characterized in that: Step S22 is specifically as follows: Step S221: collecting images of crop health status data to obtain crop health status images; Step S222: performing grayscale conversion according to the crop health status image, thereby obtaining a crop health status grayscale image; Step S223: extracting leaf texture features according to the grayscale image of the crop health status, thereby obtaining leaf texture data; Step S224: performing roughness statistics on the blade texture data to obtain high-roughness blade texture data; Step S225: dividing the crop health grayscale image into regions according to the high-roughness leaf texture data, thereby obtaining high-roughness leaf texture region data; Step S226: performing linear food mark recognition according to the leaf texture data, thereby obtaining linear food mark data; Step S227: performing an insect feeding trace intersection operation based on the linear feeding trace data and the high-roughness leaf texture area data to obtain the insect feeding trace data.

6. The intelligent collection and management method of smart agricultural information big data according to claim 5 is characterized in that: Step S226 is specifically as follows: Gray-level co-occurrence matrix calculation is performed according to leaf texture data, thereby obtaining gray-level co-occurrence matrix data; Perform entropy statistics on the gray-level co-occurrence matrix data to obtain high-entropy gray-level co-occurrence matrix data; Energy calculation is performed according to the gray-level co-occurrence matrix data, thereby obtaining gray-level co-occurrence matrix energy data; Perform energy statistics on the gray-level co-occurrence matrix energy data to obtain low-energy gray-level co-occurrence matrix data; The food trace data is obtained by performing intersection operation on the low-energy gray-level co-occurrence matrix data and the high-entropy gray-level co-occurrence matrix data; The food trace data is subjected to elongated shape recognition to obtain linear food trace data.

7. The intelligent collection and management method of smart agricultural information big data according to claim 4 is characterized in that: Step S23 is specifically as follows: Step S231: collecting images of crop physiological data to obtain crop physiological images; Step S232: filtering and denoising the crop physiological image to obtain a crop physiological denoised image; Step S233: performing wormhole edge detection according to the crop physiological denoising image, thereby obtaining wormhole edge data; Step S234: performing an expansion operation on the wormhole edge data to obtain wormhole edge expansion data; Step S235: calibrating the wormhole region of the crop physiological denoised image according to the wormhole edge expansion data, thereby obtaining wormhole region data; Step S236: performing quantity statistics according to the wormhole area data, thereby obtaining wormhole quantity data; Step S237: Calculate the area according to the wormhole area data, thereby obtaining the wormhole area data; Step S238: Perform wormhole feature fusion according to the wormhole area data and the wormhole quantity data to obtain wormhole data.

8. The intelligent collection and management method of smart agricultural information big data according to claim 1 is characterized in that: Step S3 is specifically as follows: Step S31: acquiring farm remote sensing data, and collecting crop spectra according to the farm remote sensing data, thereby obtaining crop spectral data; Step S32: dividing the crop spectral data into bands, thereby obtaining green light band data and short-wave infrared band data; Step S33: performing normalized moisture index calculation according to the green light band data and the short-wave infrared band data, thereby obtaining normalized moisture index data; Step S34: performing water stress identification according to the normalized water index data, thereby obtaining water stress data; Step S35: obtaining crop water stress threshold data; Step S36: dividing the water stress data according to the crop water stress threshold data, thereby obtaining high water state data and low water state data; Step S37: maintaining the crop irrigation plan for the high moisture state data, thereby obtaining the crop high moisture state irrigation plan data; Step S38: increasing the irrigation frequency for the low moisture state data, thereby obtaining the irrigation frequency data for the low moisture state of the crop; Step S39: Integrate the irrigation plan according to the crop high moisture state irrigation plan data and the crop low moisture state irrigation frequency data to obtain the crop irrigation plan data, and evaluate the moisture state according to the crop irrigation plan data to obtain the crop moisture state data.

9. The intelligent collection and management method of smart agricultural information big data according to claim 1 is characterized in that: Step S5 is specifically as follows: Step S51: dividing the area according to the visualized data of the area affected by pests and diseases, so as to obtain data of areas with high incidence of pests and diseases; Step S52: optimizing the pesticide spraying dosage for the pest-prone areas data, thereby obtaining the pesticide spraying dosage optimization data; Step S53: performing crop density statistics on the pest-prone area data, thereby obtaining pest-prone crop density data; Step S54: setting a density threshold according to the density data of crops with high incidence of insect pests, thereby obtaining a high incidence density threshold of insect pests; Step S55: Density monitoring is performed on the visualized data of the pest-affected area according to the pest high-incidence density threshold, so as to obtain data of potential pest high-incidence areas and data of areas that have not reached the pest density threshold; Step S56: increasing the frequency of pesticide spraying according to the data of potential high-incidence areas of insect pests, thereby obtaining pesticide spraying frequency data; Step S57: reducing the amount of pesticide spraying according to the data of the area that does not reach the pest density threshold, thereby obtaining pesticide spraying amount data; Step S58: constructing a pesticide intelligent spraying management strategy based on the pesticide spraying dosage optimization data, the pesticide spraying frequency data, and the pesticide spraying amount data, thereby obtaining pesticide spraying intelligent management data.

10. A smart agricultural information big data intelligent collection and management system, characterized in that: Used to execute the smart agricultural information big data intelligent collection and management method as claimed in claim 1, the smart agricultural information big data intelligent collection and management system comprises: A crop growth analysis module is used to obtain farm sensor data and perform crop growth analysis based on the farm sensor data to obtain crop growth data; The pest and disease detection module is used to extract crop health status characteristics and crop physiological characteristics according to crop growth data, so as to obtain crop health status data and crop physiological data; perform pest and disease detection according to crop health status data and crop physiological data, so as to obtain crop pest and disease data; A crop moisture status assessment module is used to obtain farm remote sensing data, and perform crop multispectral analysis based on the farm remote sensing data, thereby obtaining crop multispectral data, and perform crop moisture status assessment based on the crop multispectral data, thereby obtaining crop moisture status data; The pest and disease affected area visualization module is used to perform correlation analysis on crop moisture status data based on crop pest and disease data, thereby obtaining crop pest and disease-moisture status data; and to visualize the pest and disease affected area based on the crop pest and disease-moisture status data, thereby obtaining pest and disease affected area visualization data; The intelligent management module for pesticide spraying is used to perform intelligent management of pesticide spraying based on the visualized data of the area affected by pests and diseases, thereby obtaining intelligent management data for pesticide spraying.

Images