Agricultural information management system and method based on big data platform

Abstract

The present application discloses an agricultural information management system and method based on a big data platform, which relates to the field of intelligent management technology. The system acquires farmland remote sensing data collected by near-Earth satellites; extracts data features from the farmland remote sensing data to obtain a divergence topology global farmland remote sensing feature matrix; and determines the level of pests and diseases in the farmland based on the divergence topology global farmland remote sensing feature matrix. The system can timely and accurately monitor and identify the extent of pests and diseases, improve agricultural production efficiency and ensure food security.

Description

Technical Field

[0001] The present application relates to the field of intelligent management technology, and more specifically, to an agricultural information management system and method based on a big data platform. Background Art

[0002] Agriculture is an important pillar of the national economy and the foundation of human survival. However, agricultural production faces many challenges, one of which is the prevention and control of pests and diseases. Pests and diseases not only lead to reduced yields and quality of crops, but also cause ecological damage. Therefore, timely and accurate monitoring and identification of the extent of pests and diseases is the key to improving agricultural production efficiency and ensuring food security. Traditional pest and disease monitoring methods mainly rely on manual inspections and sampling analysis, which have disadvantages such as low time efficiency and incomplete spatial coverage.

[0003] With the integration and development of big data and agriculture, the construction of an agricultural big data platform has become an inevitable trend. In various links such as land planning, germplasm resource selection, pest control, production management, procurement and sales, precision marketing, storage and channel docking of agricultural products, massive data in different formats and different business fields are integrated into a standard unified data source to achieve data analysis, data mining and data visualization, timely and comprehensively grasp the development trends of agriculture, and conduct human intervention based on this, so that agriculture can develop in an ideal direction.

[0004] Therefore, a solution based on a big data platform is needed to solve the problems existing in agricultural information management. Summary of the invention

[0005] In order to solve the above technical problems, this application is proposed. The embodiment of this application provides an agricultural information management system and method based on a big data platform, which obtains farmland remote sensing data collected by low-Earth satellites; extracts data features from the farmland remote sensing data to obtain a divergence topology global farmland remote sensing feature matrix; based on the divergence topology global farmland remote sensing feature matrix, determines the level of pests and diseases in the farmland, and can timely and accurately monitor and identify the extent of pests and diseases, improve agricultural production efficiency and ensure food security.

[0006] In the first aspect, a method for agricultural information management based on a big data platform is provided, which includes:

[0007] Obtaining remote sensing data of farmland collected by near-Earth satellites;

[0008] The farmland remote sensing data include farmland remote sensing images;

[0009] Extracting data features from the farmland remote sensing data to obtain a divergence topology global farmland remote sensing feature matrix;

[0010] Based on the divergence topology global farmland remote sensing feature matrix, the pest and disease level of the farmland is determined.

[0011] In the second aspect, an agricultural information management system based on a big data platform is provided, which includes:

[0012] An image acquisition module is used to acquire remote sensing images of farmland collected by near-earth satellites;

[0013] An image feature extraction module is used to extract image features from the farmland remote sensing image to obtain a divergence topology global farmland remote sensing feature matrix;

[0014] The pest and disease level determination module is used to determine the pest and disease level of the farmland based on the divergence topology global farmland remote sensing feature matrix.

[0015] The present application has at least the following technical effects: compared with the prior art, the present application obtains farmland remote sensing data collected by low-Earth satellites, extracts data features from the farmland remote sensing data to obtain a divergence topology global farmland remote sensing feature matrix, and determines the level of pests and diseases in the farmland based on the divergence topology global farmland remote sensing feature matrix; by extracting and analyzing image features, the types of pests and diseases existing in the farmland can be judged, and compared with a known pest and disease library to further determine the level of pests and diseases; and based on the farmland information and pest and disease levels provided by satellite images, agricultural experts and decision makers can formulate corresponding prevention and control measures. The above information and analysis results can help agricultural managers better understand the situation of pests and diseases, and take corresponding measures to protect crops and improve the yield and quality of farmland. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0017] Figure 1 It is a flow chart of the agricultural information management method based on the big data platform according to the embodiment of the present application.

[0018] Figure 2 It is a schematic diagram of the architecture of the agricultural information management method based on the big data platform according to the embodiment of the present application.

[0019] Figure 3 It is a flowchart of the sub-steps of step 120 in the agricultural information management method based on the big data platform according to an embodiment of the present application.

[0020] Figure 4It is a block diagram of an agricultural information management system based on a big data platform according to an embodiment of the present application. DETAILED DESCRIPTION

[0021] The following will describe the technical solutions in the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0022] Unless otherwise specified, all technical and scientific terms used in the embodiments of the present application have the same meaning as those commonly understood by those skilled in the art of the present application. The terms used in this application are only for the purpose of describing specific embodiments and are not intended to limit the scope of this application.

[0023] In the embodiments of the present application, it should be noted that, unless otherwise specified and limited, the term "connection" should be understood in a broad sense. For example, it can be an electrical connection or a connection between two components. It can be a direct connection or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to the specific circumstances.

[0024] It should be noted that the terms "first\second\third" involved in the embodiments of the present application are only used to distinguish similar objects, and do not represent a specific order for the objects. It is understandable that the specific order or sequence of "first\second\third" can be interchanged where permitted. It should be understood that the objects distinguished by "first\second\third" can be interchanged where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein.

[0025] In agricultural production, pests and diseases have caused extensive damage to agricultural production and the ecological environment. The invasion of pests and diseases can cause crop leaves to fall, fruits to rot, and roots to be damaged, thereby reducing crop yield and quality. By timely and effective prevention and control of pests and diseases, crop losses can be reduced and crop yield and quality can be improved.

[0026] Large-scale outbreaks of pests and diseases will disrupt the balance of farmland ecosystems, destroy beneficial biological populations, increase the use of pesticides, and pollute the environment. Through scientific pest and disease control measures, we can reduce the adverse impact on the environment, maintain the ecological balance of farmland, and promote the sustainable development of agriculture.

[0027] Pest and disease control can reduce the reliance on pesticides in agricultural production, reduce the amount and cost of pesticides used. At the same time, through scientific pest and disease management measures, crop losses can be reduced, the benefits of agricultural production can be improved, production costs can be reduced, and farmers' income can be increased.

[0028] The prevention and control of pests and diseases is necessary to ensure agricultural production benefits, maintain ecological balance, reduce production costs and ensure food security. Through the application of science and technology and innovative methods, we can improve the monitoring and identification capabilities of pests and diseases, formulate scientific prevention and control strategies, and achieve sustainable agricultural development and food security.

[0029] In this application, a method for agricultural information management based on a big data platform is provided, which can realize data analysis, data mining and data visualization, timely and comprehensively grasp the development trends of agriculture, and can carry out human intervention based on this, so that agriculture can develop in an ideal direction.

[0030] Figure 1 It is a flow chart of the agricultural information management method based on the big data platform according to the embodiment of the present application. Figure 2 FIG. 1 is a schematic diagram of the architecture of the agricultural information management method based on the big data platform according to the embodiment of the present application. Figure 1 and Figure 2 As shown, the agricultural information management method based on the big data platform includes: step 110, acquiring a farmland remote sensing image collected by a near-earth satellite; step 120, performing image feature extraction on the farmland remote sensing image to obtain a divergence topology global farmland remote sensing feature matrix; step 130, determining the pest and disease level of the farmland based on the divergence topology global farmland remote sensing feature matrix.

[0031] Wherein, in the step 110, the image quality and resolution of the satellite acquisition are ensured to obtain clear and accurate remote sensing images of farmland. The temporal and spatial coverage of the image is ensured to obtain comprehensive farmland information. Different satellite data sources and image acquisition strategies may be required for different farmland types and geographical locations. Wherein, farmland remote sensing images can provide large-scale, high-resolution farmland information, including vegetation growth status, soil moisture, temperature distribution, etc. The spatial distribution and dynamic change information of farmland can be obtained in a timely manner to provide basic data for pest and disease control. Through satellite images, rapid monitoring and evaluation of large areas of farmland can be achieved, and the efficiency of pest and disease monitoring can be improved.

[0032] In step 120, a suitable image feature extraction algorithm is selected, such as texture features, shape features, color features, etc., to capture information related to pests and diseases in farmland remote sensing images. Different feature extraction methods may be used for different types of pests and diseases to obtain more accurate feature representations. Among them, image feature extraction can convert farmland remote sensing images into numerical feature representations, which is convenient for subsequent data processing and analysis. The divergence topology global farmland remote sensing feature matrix can comprehensively reflect the overall characteristics of the farmland, including vegetation conditions, soil texture, water distribution, etc., providing a basis for determining the level of pests and diseases.

[0033] In step 130, a suitable pest and disease grade classification standard is established, taking into account the severity of different pest and disease types and their impact on crops. In combination with historical data and expert knowledge, the feature matrix is ​​analyzed and modeled to determine the pest and disease grade. Among them, by determining the pest and disease grade, the pest and disease problems existing in the farmland can be identified in time, and accurate prevention and control suggestions can be provided to farmers. Based on the pest and disease grade, targeted prevention and control measures can be implemented to reduce the use of pesticides, improve prevention and control effects, and reduce production costs. The distribution and development trend of farmland pests and diseases can be predicted and monitored, providing decision makers with a reference for agricultural production management.

[0034] Agricultural information management methods based on big data platforms can improve the efficiency and accuracy of pest and disease monitoring and provide scientific and intelligent support for agricultural production.

[0035] Specifically, in step 110, a remote sensing image of farmland collected by a near-earth satellite is obtained. In view of the above technical problems, the technical concept of this application is to use satellite equipment and image recognition technology to perform high-resolution remote sensing monitoring of farmland, and then automatically identify the level of pests and diseases.

[0036] Based on this, in the technical solution of this application, first, remote sensing images of farmland collected by low-Earth satellites are obtained. It is worth mentioning that images collected by low-Earth satellites can cover a large range of farmland areas. Compared with traditional pest inspection methods, such as manual inspections and sampling analysis, satellite images can simultaneously monitor and analyze large areas of farmland, providing more comprehensive pest monitoring data.

[0037] In this application, useful information that can be extracted from farmland remote sensing images is used to finally determine the level of pests and diseases in the farmland. It includes: Vegetation index: Vegetation index, such as normalized difference vegetation index (NDVI), can be calculated through farmland remote sensing images. Vegetation index can reflect the status of farmland vegetation, including the coverage, growth status and health status of vegetation. Pests and diseases usually cause vegetation withering, leaf discoloration and other phenomena. The presence and extent of pests and diseases can be preliminarily judged by the changes in vegetation index.

[0038] Texture features: Texture features in farmland remote sensing images can be used to identify and classify pests and diseases. Different pests and diseases will cause different texture changes in farmland images, such as spots, patches, stripes, etc. By extracting and analyzing the texture features of the image, the type of pests and diseases in the farmland can be determined, and the level of pests and diseases can be further determined.

[0039] Shape features: Shape features in farmland remote sensing images can provide information about farmland structure and crop morphology. Pests and diseases usually cause morphological changes in crops, such as leaf deformities and fruit deformations. By extracting and analyzing shape features in images, we can determine the types of pests and diseases in the farmland and assess their impact on crops.

[0040] Color features: The color features in farmland remote sensing images can reflect the growth status of crops and the presence of pests and diseases. Different pests and diseases can cause changes in the color of crop leaves, such as yellowing and withering. By extracting and analyzing the color features in the image, the presence and extent of pests and diseases can be preliminarily determined.

[0041] Spatial distribution information: Remote sensing images of farmland provide spatial distribution information of farmland, which can help determine the scope and distribution of pests and diseases. By analyzing the spatial distribution of pests and diseases in the images, the impact of pests and diseases on farmland can be assessed, and the pest and disease grade can be further determined.

[0042] Useful information that can be extracted from farmland remote sensing images includes vegetation index, texture features, shape features, color features and spatial distribution information. By extracting and analyzing this information, we can preliminarily judge the existence and extent of pests and diseases, and ultimately determine the pest and disease level of the farmland. This information is of great significance for formulating corresponding prevention and control measures and protecting crops.

[0043] Specifically, remote sensing images of farmland collected by satellites can provide a wide range of farmland information, including vegetation conditions, soil moisture, temperature distribution, etc. This information can help agricultural experts and decision makers understand the overall status of farmland and provide a basis for determining the level of pests and diseases. Satellite images have the characteristics of wide coverage and high resolution, which can realize rapid monitoring of large areas of farmland. By analyzing remote sensing images of farmland, the occurrence and spread of pests and diseases can be discovered in time, helping agricultural managers to grasp the overall situation of pests and diseases. Satellite images can provide time-series change information of farmland, including vegetation growth process, development trend of pests and diseases, etc. By comparing and analyzing satellite images of multiple periods, we can understand the evolution process of pests and diseases, judge the severity and scope of impact of pests and diseases, and thus determine the level of pests and diseases.

[0044] Image features in farmland remote sensing images can be used to identify and classify pests and diseases. By extracting and analyzing image features, the types of pests and diseases in the farmland can be determined, and compared with the known pest and disease database to further determine the pest and disease level. Based on the farmland information and pest and disease levels provided by satellite images, agricultural experts and decision makers can formulate corresponding prevention and control measures. Depending on the level of pests and diseases, different levels of prevention and control measures can be taken, including adjusting farmland management measures, using appropriate pesticides and biological control methods, etc., to minimize the losses caused by pests and diseases.

[0045] Acquiring remote sensing images of farmland collected by near-Earth satellites plays an important role in determining the pest and disease level of farmland. It can provide comprehensive farmland information, achieve large-scale monitoring, provide time series change information, assist in the identification and classification of pests and diseases, and support decision-making and the formulation of prevention and control measures. This information and analysis results can help agricultural managers better understand the situation of pests and diseases, take appropriate measures to protect crops and improve the yield and quality of farmland.

[0046] Specifically, in step 120, image features are extracted from the farmland remote sensing image to obtain a divergence topology global farmland remote sensing feature matrix. Figure 3 FIG. 1 is a flowchart of the sub-steps of step 120 in the agricultural information management method based on a big data platform according to an embodiment of the present application. Figure 3 As shown, image feature extraction is performed on the farmland remote sensing image to obtain a divergence topology global farmland remote sensing feature matrix, including: step 121, image block processing is performed on the farmland remote sensing image to obtain a sequence of farmland remote sensing image blocks; step 122, global correlation feature information is extracted from the sequence of farmland remote sensing image blocks to obtain the divergence topology global farmland remote sensing feature matrix.

[0047] Furthermore, the farmland remote sensing image is processed by block processing, and the image is divided into multiple blocks, each block represents a small area. In this way, the local information of the image can be extracted, which is convenient for subsequent feature extraction and analysis. The block processing can be implemented using a sliding window or other block algorithms.

[0048] Extract global correlation feature information from the sequence of farmland remote sensing image blocks to obtain the divergence topology global farmland remote sensing feature matrix. These features may include:

[0049] Vegetation index statistical characteristics: Calculate the statistical characteristics of the vegetation index of the entire image block sequence, such as mean, variance, maximum, minimum, etc. These characteristics can reflect the overall vegetation status of the farmland.

[0050] Texture feature extraction: Extract texture features from each block in the image block sequence, such as gray level co-occurrence matrix (GLCM) features, local binary pattern (LBP) features, etc. Then, the statistical information of the texture features of the entire image block sequence can be calculated, such as mean, variance, energy, entropy, etc. These features can reflect the texture changes of the farmland.

[0051] Shape feature extraction: Extract shape features from each block in the image block sequence, such as area, perimeter, compactness, etc. Then, the statistical information of the shape features of the entire image block sequence can be calculated, such as the mean, variance, maximum, minimum, etc. These features can reflect the morphological changes of the farmland.

[0052] Color feature extraction: Extract color features from each block in the image block sequence, such as color histogram, color moment, etc. Then, the statistical information of the color features of the entire image block sequence can be calculated, such as the mean, variance, maximum, minimum, etc. These features can reflect the color changes of the farmland.

[0053] Spatial distribution feature extraction: By performing spatial distribution analysis on each block in the image block sequence, we can calculate the relative position, distance, and distribution of the blocks. These features can reflect the spatial distribution law of farmland.

[0054] By extracting the above global correlation feature information, the divergence topology global farmland remote sensing feature matrix can be obtained. This feature matrix contains the global feature information of farmland remote sensing images and can be used for further pest and disease grade judgment and decision support. The application of this feature matrix helps to improve the effect and accuracy of farmland remote sensing image analysis.

[0055] For step 121, then, the farmland remote sensing image is subjected to image block processing to obtain a sequence of farmland remote sensing image blocks; and global correlation feature information is extracted from the sequence of farmland remote sensing image blocks to obtain the divergence topology global farmland remote sensing feature matrix. That is, the overall distribution and correlation of farmland pests and diseases are captured from the sequence of farmland remote sensing image blocks. Specifically, by extracting the global correlation features of the sequence of farmland remote sensing image blocks, the overall distribution of farmland pests and diseases can be obtained with a global perspective. This global perspective can better understand the propagation pattern and diffusion trend of pests and diseases in farmland.

[0056] Image segmentation is the process of dividing an image into multiple blocks or regions so that each block can be processed and analyzed independently. Image segmentation of farmland remote sensing images can help extract local features and perform more detailed analysis.

[0057] Specifically, the size of each image block needs to be determined first. The selection of the block size should be adjusted according to the specific application requirements and image characteristics. Generally, the size of the block can be a fixed square or rectangular area, or it can be adaptively adjusted according to the image content. Then, the position of each image block can be determined. The sliding window method can be used to slide the window on the image with a certain step size, and the area covered by the window is taken as an image block. The choice of step size determines the degree of overlap between blocks. A smaller step size can increase the overlap between blocks and provide more information. Then, according to the determined block size and position, the corresponding image block is extracted from the original image. The image block extraction operation can be implemented using an image processing library or software. The extracted image blocks can be saved as a sequence for subsequent feature extraction and analysis. Finally, each image block is subjected to subsequent feature extraction, analysis or other operations. Specific processing can be performed on each image block, or the results of all image blocks can be integrated.

[0058] It should be noted that image block processing may result in loss or duplication of information, so a trade-off needs to be made when selecting the block size and step size. In addition, the results of block processing may be affected by the edge effect of the image, which can be reduced by adding borders around the edge of the image or using appropriate border processing methods.

[0059] That is, image block processing is the process of dividing farmland remote sensing images into multiple blocks or regions so that each block can be processed and analyzed independently. This can extract local features and perform more refined farmland remote sensing image analysis.

[0060] The step 122 includes: extracting node feature information and topological feature information from the sequence of farmland remote sensing image blocks respectively to obtain a sequence of farmland remote sensing feature vectors and a divergence value topological feature matrix; and passing the sequence of farmland remote sensing feature vectors and the divergence value topological feature matrix through a graph neural network model to obtain the divergence topology global farmland remote sensing feature matrix.

[0061] In one embodiment of the present application, node feature information and topological feature information are respectively extracted from the sequence of the farmland remote sensing image blocks to obtain a sequence of farmland remote sensing feature vectors and a divergence value topological feature matrix, including: passing the sequence of the farmland remote sensing image blocks through a remote sensing image feature extractor based on a first convolutional neural network model to obtain a sequence of the farmland remote sensing feature vectors; calculating the JS divergence value between any two farmland remote sensing feature vectors in the sequence of the farmland remote sensing feature vectors to obtain a divergence value topological matrix; and, passing the divergence value topological matrix through a divergence topological feature extractor based on a second convolutional neural network model to obtain the divergence value topological feature matrix.

[0062] In a specific example of the present application, the encoding process of extracting global correlation feature information from the sequence of farmland remote sensing image blocks to obtain the divergence topology global farmland remote sensing feature matrix includes: first, passing the sequence of farmland remote sensing image blocks through a remote sensing image feature extractor based on a first convolutional neural network model to obtain a sequence of farmland remote sensing feature vectors; at the same time, calculating the JS divergence value between any two farmland remote sensing feature vectors in the sequence of farmland remote sensing feature vectors to obtain a divergence value topology matrix; then, passing the divergence value topology matrix through a divergence topology feature extractor based on a second convolutional neural network model to obtain a divergence value topology feature matrix; and then passing the sequence of farmland remote sensing feature vectors and the divergence value topology feature matrix through a graph neural network model to obtain a divergence topology global farmland remote sensing feature matrix.

[0063] By extracting the sequence feature vectors of farmland remote sensing image blocks through the remote sensing image feature extractor based on the first convolutional neural network model, the feature representation of each image block can be obtained. These feature vectors can capture the key information in the farmland remote sensing image. Next, the JS divergence value between any two farmland remote sensing feature vectors is calculated to obtain the divergence value topological matrix, which can reflect the similarities and differences between different image blocks.

[0064] Then, the divergence value topology matrix is ​​processed by a divergence topology feature extractor based on the second convolutional neural network model to extract a divergence value topology feature matrix. The divergence value topology feature matrix can capture the structural information of the divergence value topology and reflect the topological relationship between image blocks.

[0065] Finally, the sequence of farmland remote sensing feature vectors and the divergence value topological feature matrix are integrated through the graph neural network model to obtain the divergence topological global farmland remote sensing feature matrix. This feature matrix integrates the feature representation of image blocks, divergence value topological features and the relationship between them, and can provide more comprehensive farmland remote sensing information.

[0066] Specifically, the advantages of convolutional neural networks, divergence value topology and graph neural networks can be combined to capture the characteristics of farmland remote sensing images from different angles and provide a more global feature matrix. This feature matrix can be used for pest and disease grade judgment and decision support, which helps to improve the accuracy and effect of farmland remote sensing image analysis.

[0067] Specifically, in step 130, the pest and disease level of the farmland is determined based on the divergence topology global farmland remote sensing feature matrix, including: passing the divergence topology global farmland remote sensing feature matrix through a classifier to obtain a classification result, and the classification result is used to represent the pest and disease level label of the farmland.

[0068] Next, the feature matrix can be input into the classifier for classification. The classifier can use various machine learning algorithms, such as support vector machine (SVM), random forest, neural network, etc. The classifier will learn the patterns between different pest and disease levels based on the input feature matrix and map it to the corresponding classification results, that is, the pest and disease level label of the farmland.

[0069] By inputting the divergence topology global farmland remote sensing feature matrix into the classifier, the classification result of the farmland pest level can be obtained. This classification result can be used to judge the pest level of the farmland and provide useful information for farmland management and decision-making.

[0070] In the technical solution of the present application, when the sequence of the farmland remote sensing feature vectors and the divergence value topological feature matrix are passed through a graph neural network model to obtain a divergence topological global farmland remote sensing feature matrix, each row feature vector of the divergence topological global farmland remote sensing feature matrix expresses the topological association expression of the image semantic features of the farmland remote sensing image under the JS divergence value topological relationship. Therefore, considering the relative independence of the topological expression of each row feature vector based on the graph neural network model, the divergence topological global farmland remote sensing feature matrix as a feature distribution as a whole has local-global fine-grained semantic graph coding aggregation sparsity, and therefore it is expected to improve the graph coding fine-grained semantic coding aggregation expression effect of the divergence topological global farmland remote sensing feature matrix.

[0071] In a preferred embodiment, the divergence topology global farmland remote sensing feature matrix is ​​passed through a classifier to obtain a classification result, and the classification result is used to represent the pest and disease grade label of the farmland.

[0072] Calculate the absolute value sum of each eigenvalue of the divergence topology global farmland remote sensing feature matrix to obtain a first divergence topology global farmland remote sensing and modulation value, and calculate the square root of the sum of the squares of each eigenvalue of the divergence topology global farmland remote sensing feature matrix to obtain a second divergence topology global farmland remote sensing and modulation value;

[0073] After performing point subtraction on the divergence topology global farmland remote sensing feature matrix and the second divergence topology global farmland remote sensing and modulation value, respectively perform point multiplication with the number of eigenvalues ​​of the divergence topology global farmland remote sensing feature matrix and the reciprocal of the first divergence topology global farmland remote sensing and modulation value, and take the reciprocal of each eigenvalue to obtain the first divergence topology global farmland remote sensing phase conversion matrix;

[0074] After point subtraction of the divergence topology global farmland remote sensing feature matrix and the first divergence topology global farmland remote sensing and modulation value, point multiplication is performed with the square root of the number of eigenvalues ​​of the divergence topology global farmland remote sensing feature matrix and the reciprocal of the second divergence topology global farmland remote sensing and modulation value, and the reciprocal of each eigenvalue is taken to obtain the second divergence topology global farmland remote sensing phase conversion matrix;

[0075] Subtract the dot product matrix of the second divergence topology global farmland remote sensing phase conversion matrix and the weighted hyperparameter from the first divergence topology global farmland remote sensing phase conversion matrix to obtain an optimized divergence topology global farmland remote sensing feature matrix;

[0076] The optimized divergence topology global farmland remote sensing feature matrix is ​​passed through a classifier to obtain a classification result.

[0077] Among them, the optimization expression of the divergence topology global farmland remote sensing feature matrix M is:

[0078]

[0079]

[0080] m i ∈M∈R W×H

[0081] n=W×H

[0082] Wherein, M is the global farmland remote sensing feature matrix of the divergence topology, R represents a real number set, W and H represent the number of rows and columns of the global farmland remote sensing feature matrix of the divergence topology, respectively, and m i represents the eigenvalue of the i-th position of the divergence topology global farmland remote sensing feature matrix, n represents the number of eigenvalues ​​of the divergence topology global farmland remote sensing feature matrix, α represents the first divergence topology global farmland remote sensing and modulation value, β represents the second divergence topology global farmland remote sensing and modulation value, ⊙ represents the point multiplication by position, Indicates positional subtraction, (·) ⊙-1 Represents the inverse of each eigenvalue of the calculated feature matrix, M1 represents the first divergence topology global farmland remote sensing phase conversion matrix, M2 represents the second divergence topology global farmland remote sensing phase conversion matrix, ω represents the weighted hyperparameter, and M′ represents the optimized divergence topology global farmland remote sensing feature matrix.

[0083] Therefore, in the preferred example, the difference of the difference and modulation representation of the eigenvalue of the divergence topology global farmland remote sensing feature matrix relative to the overall feature set of the divergence topology global farmland remote sensing feature matrix is ​​used as the semantic change intensity information, and a phase-like transformation corresponding to the position-based intensity modulation is performed through different and modulation representations, so that the aggregation enhancement of semantic change phase perception can improve the axial aggregation receptive field along the feature aggregation direction by performing a spatial translation operation based on alternating stacking under the set scale balance of the divergence topology global farmland remote sensing feature matrix, thereby improving the perception effect of the aggregation semantics of the divergence topology global farmland remote sensing feature matrix on the detailed semantic changes, so as to improve the expression effect of the divergence topology global farmland remote sensing feature matrix and improve the accuracy of the classification results obtained by passing it through the classifier.

[0084] In summary, the agricultural information management method 100 based on the big data platform according to the embodiment of the present application is explained, which uses satellite equipment and image recognition technology to perform high-resolution remote sensing monitoring of farmland to automatically identify the level of pests and diseases.

[0085] In one embodiment of the present application, Figure 4 FIG. 1 is a block diagram of an agricultural information management system based on a big data platform according to an embodiment of the present application. Figure 4 As shown, according to the agricultural information management system 200 based on the big data platform according to the embodiment of the present application, it includes: an image acquisition module 210, which is used to acquire farmland remote sensing images collected by low-Earth satellites; an image feature extraction module 220, which is used to extract image features of the farmland remote sensing images to obtain a divergence topology global farmland remote sensing feature matrix; and a pest and disease level determination module 230, which is used to determine the pest and disease level of the farmland based on the divergence topology global farmland remote sensing feature matrix.

[0086] Specifically, in the agricultural information management system based on the big data platform, the image feature extraction module includes: an image block processing unit, which is used to perform image block processing on the farmland remote sensing image to obtain a sequence of farmland remote sensing image blocks; a global correlation feature extraction unit, which is used to extract global correlation feature information from the sequence of farmland remote sensing image blocks to obtain the divergence topology global farmland remote sensing feature matrix.

[0087] Specifically, in the agricultural information management system based on the big data platform, the global correlation feature extraction unit is used to: extract node feature information and topological feature information from the sequence of the farmland remote sensing image blocks respectively to obtain a sequence of farmland remote sensing feature vectors and a divergence value topological feature matrix; and, pass the sequence of farmland remote sensing feature vectors and the divergence value topological feature matrix through a graph neural network model to obtain the divergence topology global farmland remote sensing feature matrix.

[0088] Among them, those skilled in the art can understand that the specific functions and operations of each unit and module in the above-mentioned agricultural information management system based on the big data platform have been referred to above. Figures 1 to 3 The description of the agricultural information management method based on big data platform has been introduced in detail, and therefore, its repeated description will be omitted.

[0089] As described above, the agricultural information management system 200 based on the big data platform according to the embodiment of the present application can be implemented in various terminal devices, such as a server for agricultural information management based on the big data platform. In one example, the agricultural information management system 200 based on the big data platform according to the embodiment of the present application can be integrated into the terminal device as a software module and / or a hardware module. For example, the agricultural information management system 200 based on the big data platform can be a software module in the operating system of the terminal device, or it can be an application developed for the terminal device; of course, the agricultural information management system 200 based on the big data platform can also be one of the many hardware modules of the terminal device.

[0090] Alternatively, in another example, the agricultural information management system 200 based on the big data platform and the terminal device may also be separate devices, and the agricultural information management system 200 based on the big data platform may be connected to the terminal device via a wired and / or wireless network, and transmit interactive information in accordance with an agreed data format.

[0091] It should also be noted that in the apparatus, device and method of the present application, each component or each step can be decomposed and / or recombined. Such decomposition and / or recombination should be regarded as equivalent solutions of the present application.

[0092] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present application. Therefore, the present application is not intended to be limited to the aspects shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

[0093] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present application to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims

1. The agricultural information management method based on the big data platform is characterized by: include: Obtaining remote sensing data of farmland collected by near-Earth satellites; Data feature extraction is performed on the farmland remote sensing data to obtain a divergence topology global farmland remote sensing feature matrix, which includes: Performing data block processing on the farmland remote sensing data to obtain a sequence of farmland remote sensing data blocks; Extracting global correlation feature information from the sequence of farmland remote sensing data blocks to obtain the divergence topology global farmland remote sensing feature matrix, which includes: Node feature information and topological feature information are respectively extracted from the sequence of farmland remote sensing data blocks to obtain a sequence of farmland remote sensing feature vectors and a divergence value topological feature matrix, which includes: Passing the sequence of farmland remote sensing data blocks through a remote sensing data feature extractor based on a first convolutional neural network model to obtain a sequence of farmland remote sensing feature vectors; Calculating the JS divergence value between any two farmland remote sensing feature vectors in the sequence of farmland remote sensing feature vectors to obtain a divergence value topological matrix; Passing the divergence value topological matrix through a divergence topological feature extractor based on a second convolutional neural network model to obtain the divergence value topological feature matrix; The sequence of the farmland remote sensing feature vectors and the divergence value topological feature matrix are passed through a graph neural network model to obtain the divergence topological global farmland remote sensing feature matrix; Based on the divergence topology global farmland remote sensing feature matrix, the pest and disease level of the farmland is determined.

2. The agricultural information management method based on the big data platform according to claim 1 is characterized in that: Based on the divergence topology global farmland remote sensing feature matrix, the pest and disease level of the farmland is determined, including: The divergence topology global farmland remote sensing feature matrix is ​​passed through a classifier to obtain a classification result, and the classification result is used to represent the pest and disease grade label of the farmland.

3. The agricultural information management system based on the big data platform is characterized by: include: An image acquisition module is used to acquire remote sensing images of farmland collected by near-earth satellites; An image feature extraction module is used to extract image features from the farmland remote sensing image to obtain a divergence topology global farmland remote sensing feature matrix, which includes: An image block processing unit, used for performing image block processing on the farmland remote sensing image to obtain a sequence of farmland remote sensing image blocks; A global correlation feature extraction unit is used to extract global correlation feature information from the sequence of farmland remote sensing image blocks to obtain the divergence topology global farmland remote sensing feature matrix, which includes: Node feature information and topological feature information are respectively extracted from the sequence of farmland remote sensing image blocks to obtain a sequence of farmland remote sensing feature vectors and a divergence value topological feature matrix, which includes: Passing the sequence of farmland remote sensing data blocks through a remote sensing data feature extractor based on a first convolutional neural network model to obtain a sequence of farmland remote sensing feature vectors; Calculating the JS divergence value between any two farmland remote sensing feature vectors in the sequence of farmland remote sensing feature vectors to obtain a divergence value topological matrix; Passing the divergence value topological matrix through a divergence topological feature extractor based on a second convolutional neural network model to obtain the divergence value topological feature matrix; The sequence of the farmland remote sensing feature vectors and the divergence value topological feature matrix are passed through a graph neural network model to obtain the divergence topological global farmland remote sensing feature matrix; The pest and disease level determination module is used to determine the pest and disease level of the farmland based on the divergence topology global farmland remote sensing feature matrix.

Images