A method for selecting ground investigation points of crop growth remote sensing inversion products
Patent Information
- Application Number
- CN202611015093.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-09
- Publication Date
- 2026-10-27
AI Technical Summary
例如，在均质的大田区域和异质的丘陵区域采用相同密度布设地面调查点，会造成资源浪费或关键信息缺失
[0013]本发明的优点在于：本发明所选取地面调查点能更好代表作物种植区内各种可能的“长势-环境”组合类型。用此方法选取的地面调查点对作物长势遥感反演产品进行验证，能够更准确地评估作物长势遥感反演产品在不同条件下的反演精度。在保证代表性的前提下，引入基于道路网络的可达性分析，可显著降低实地调查的时间、人力和经济成本。
Smart Images
Figure CN122990458A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to a method for selecting ground survey points for crop growth remote sensing inversion products, belonging to the field of agricultural remote sensing technology. Background Technology
[0002] Monitoring crop growth over large areas using remote sensing inversion products (such as crop leaf area index, crop canopy chlorophyll content, and crop canopy nitrogen content) has become an important tool in modern agricultural management. However, due to varying degrees of error in data acquisition, radiometric calibration, geometric correction, atmospheric correction, and inversion, crop growth remote sensing inversion products rely on verification and correction using ground survey data. Therefore, scientifically and efficiently selecting representative ground survey points is crucial for better utilizing the capabilities of crop growth remote sensing inversion products.
[0003] Currently, commonly used methods for selecting ground survey points include random sampling and systematic sampling. Random sampling involves distributing survey points completely randomly within the crop planting area. While simple in statistical terms and easy to implement, the resulting survey points are often geographically unevenly distributed, leading to insufficient representativeness. Systematic sampling distributes survey points at fixed spatial intervals. This method provides uniform coverage of the crop planting area, but it does not consider the spatial heterogeneity of crop growth. For example, using the same density of survey points in homogeneous field areas and heterogeneous hilly areas can result in resource waste or missing crucial information. Furthermore, these methods do not consider environmental factors affecting crop growth, such as topography, soil conditions, management practices, and meteorological factors, resulting in survey points that cannot comprehensively cover crop growth types under different environmental conditions. They also fail to consider constraining spatial information such as road accessibility and patch boundaries, potentially leading to difficulties in conducting on-site surveys at some points.
[0004] Therefore, there is an urgent need to develop a ground survey point selection method that comprehensively considers environmental heterogeneity and constrained spatial information in order to improve the scientificity and efficiency of crop growth remote sensing monitoring product verification and correction. Summary of the Invention
[0005] The purpose of this invention is to provide an objective and efficient method for selecting ground survey points for crop growth remote sensing inversion products.
[0006] To achieve the above objectives, the present invention provides the following technical solution: A method for selecting ground survey points for crop growth remote sensing inversion products includes the following steps: Obtain remote sensing inversion products of crop growth in the area to be investigated; Collect and process environmental factor data that affect the spatial differentiation of crop growth, and construct an environmental factor raster dataset; The crop growth remote sensing inversion product is overlaid with the environmental factor raster dataset, and all pixels in the crop planting area are clustered to obtain the candidate regions after clustering. Within each candidate region, patches with an area greater than a preset threshold are identified and retained as retained patches. Edge pixels of the retained patches are removed to obtain core patches. Identify candidate ground survey points for each core patch; For each candidate region, its core patches are sorted, and based on the sorting results, the candidate ground survey points of the top N core patches are selected as the final ground survey points.
[0007] Preferably, the crop growth remote sensing inversion product includes crop leaf area index and / or crop canopy chlorophyll content and / or crop canopy nitrogen content and / or crop canopy carotenoid content.
[0008] Preferably, the environmental factor raster dataset includes: topographic factors and / or soil factors and / or management factors and / or meteorological factors; The terrain factors include: elevation, slope, aspect, and terrain relief raster data; The soil factors include: raster data of soil type, soil texture, and soil organic matter content; The management factors include: irrigation / dryland zoning raster data; The meteorological factors include: temperature, precipitation, and sunshine duration raster data.
[0009] Preferably, the edge pixel removal is carried out in the following way: performing one or two rounds of pixel removal where there are no neighbors on one side.
[0010] Preferably, the method for determining the candidate ground survey points corresponding to each core patch is as follows: Determine the maximum inscribed rectangle, maximum inscribed square, or maximum inner circle for each core patch, and use the center of the maximum inscribed rectangle, maximum inscribed square, or maximum inner circle of the core patch as the candidate ground survey point for the corresponding core patch.
[0011] Preferably, determining the maximum inscribed rectangle, maximum inscribed square, or maximum inner circle for each core patch includes: The core patch is converted into a vector polygon, and the boundary of the vector polygon is converted into a graph consisting of nodes and edges, wherein the features of the nodes include their coordinates. The graph is embedded into a multidimensional feature vector, and this multidimensional feature vector is used as the embedding vector of the graph; The vertex coordinates of the largest inscribed rectangle or square, or the center coordinates and radius of the largest inner circle, are obtained from the embedding vector of the graph.
[0012] Preferably, the specific method for sorting the core patches is as follows: For each candidate area, it is sorted from largest to smallest based on the area of its core patch, then sorted from largest to smallest based on the area of the largest inscribed rectangle, largest inscribed square, or largest inner circle of the core patch, and finally sorted from smallest to largest based on the minimum distance from the candidate ground survey point corresponding to the core patch to the road network.
[0013] The advantages of this invention are: the ground survey points selected by this invention can better represent various possible combinations of "growth-environment" within the crop planting area. Using these ground survey points to validate remote sensing inversion products of crop growth can more accurately assess the inversion accuracy of the products under different conditions. While ensuring representativeness, the introduction of accessibility analysis based on road networks can significantly reduce the time, manpower, and economic costs of field surveys. Attached Figure Description
[0014] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof.
[0015] Figure 1 This is a schematic diagram of the method flow of the present invention. Detailed Implementation
[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0017] Example 1 like Figure 1 As shown, a method for selecting ground survey points for crop growth remote sensing inversion products includes the following steps: S1: Obtain remote sensing inversion products of crop growth in the area to be investigated; S2: Collect and process environmental factor data that affect the spatial differentiation of crop growth, and construct an environmental factor raster dataset; S3: Overlay the crop growth remote sensing inversion product with the environmental factor raster dataset, and cluster all pixels in the crop planting area to obtain the clustered candidate area; S4: Within each candidate region, identify and retain patches with an area greater than a preset threshold as retained patches, and remove edge pixels from the retained patches to obtain core patches; S5: Determine the candidate ground survey points corresponding to each core patch; S6: For each candidate region, sort its core patches, and select the top N core patches as the final ground survey points based on the sorting results.
[0018] As a refinement of the above embodiments, the area to be investigated in this embodiment is a county, the crops include wheat, corn, rice, etc., and the crop growth remote sensing inversion products include one or a combination of crop leaf area index (LAI), crop canopy chlorophyll content, crop canopy nitrogen content, crop canopy carotenoid content, etc.
[0019] As a refinement of the above embodiments, step S2 specifically includes: Collect and process environmental factor data that affect the spatial differentiation of crop growth, and form an environmental factor raster dataset with a unified spatial resolution and projection coordinate system. Ensure that the spatial resolution and projection coordinate system of the environmental factor raster dataset are consistent with the spatial resolution and projection coordinate system of the crop growth remote sensing inversion product.
[0020] The environmental factors include one or a combination of the following factors: Topographic factors: elevation, slope, aspect, topographic relief raster data, etc.
[0021] Soil factors: soil type, soil texture, soil organic matter content raster data, etc.
[0022] Management factors: Irrigation / dryland zoning raster data, etc.
[0023] Meteorological factors: temperature, precipitation, sunshine duration raster data, etc.
[0024] As a refinement of the above embodiments, step S3 limits all analysis to the corresponding crop planting area; specifically, it includes: The acquired remote sensing inversion products of crop growth were overlaid with the constructed environmental factor raster dataset, and clustering algorithms (such as K-Prototypes) were used to perform cluster analysis on all pixels within the crop planting area. After clustering, the crop planting area was divided into several candidate regions that were internally homogeneous but mutually heterogeneous, and each candidate region included several patches.
[0025] As a refinement of the above embodiments, step S4 includes S401: Within each candidate region, spatial filtering and connectivity analysis are performed on the clustering results to remove fragmented patches and retain patches with an area greater than a preset threshold, thus obtaining retained patches.
[0026] S402: Remove edge pixels of the retained patches according to the specified rule 1 to obtain the core patches, and calculate and record the area of the core patches.
[0027] Optionally, rule 1 may involve removing pixels with no neighbors on one side in one round, or removing pixels with no neighbors on one side in two rounds.
[0028] As a refinement of the above embodiment, step S5 determines the candidate ground survey points corresponding to each core patch and records the coordinates and other relevant information of the candidate ground survey points corresponding to each core patch.
[0029] Specifically, the method for determining the candidate ground survey points for each core patch is as follows: S501: First determine the maximum inscribed rectangle, maximum inscribed square, or maximum inner circle of each core patch, and use the center of the maximum inscribed rectangle, maximum inscribed square, or maximum inner circle of the core patch as the candidate ground survey point for the corresponding core patch. S502: Calculate the minimum distance from the candidate ground survey point to the road network, and record the coordinates of the candidate ground survey point corresponding to each core patch, its minimum distance to the road network, and the corresponding road name. At the same time, record the area of the largest inscribed rectangle, the largest inscribed square, or the largest inner circle of the core patch, the candidate area to which the candidate ground survey point belongs, the value of the crop growth remote sensing inversion product at its location, the environmental factor value, and other information.
[0030] Optionally, the road network is a road network consisting of national highways, provincial highways, county roads, and township roads.
[0031] Optionally, determining the maximum inscribed rectangle, maximum inscribed square, or maximum inner circle of each core patch includes: (1) Convert the core patch into a vector polygon, and treat the boundary of the vector polygon as, simplified or converted into a graph composed of nodes and edges, wherein the features of the nodes include their coordinates; (2) Embed the graph as a multidimensional feature vector and use it as the embedding vector of the graph; use a neural network model to obtain the vertex coordinates of the largest inscribed rectangle or the largest inscribed square or the center coordinates and radius of the largest inner circle according to the embedding vector of the graph.
[0032] In specific implementation, the above steps involve embedding the graph into a multidimensional feature vector, which may specifically include: using graph embedding technology to embed the graph into a multidimensional feature vector and using it as the embedding vector of the graph.
[0033] As a refinement of the above embodiments, step S6 specifically includes: S601: For each candidate region, sort the core patches according to the specified rule 2.
[0034] Optionally, rule 2 is to first sort the core patches by area from largest to smallest, then sort them by area of the largest inscribed rectangle, largest inscribed square, or largest inner circle of the core patches from largest to smallest, and finally sort them by minimum distance from the candidate ground survey point corresponding to the core patch to the road network from smallest to largest.
[0035] S602: For each candidate area, select ground survey points according to the specified rule 3. Optionally, rule 3 is: based on the sorting result of S601, select the candidate ground survey points of the top N core patches as the final ground survey points, where N≥1.
[0036] S603: Output ground survey points: Output the geographic coordinates (e.g., longitude and latitude) of each final ground survey point, its minimum distance to the road network, and the corresponding road name. Also output the candidate area to which each final ground survey point belongs, the value of the crop growth remote sensing inversion product at its location, environmental factor values, and the area of the largest inscribed rectangle, largest inscribed square, or largest inner circle of the corresponding core patch, facilitating subsequent field surveys and data analysis.
[0037] It should be noted that the ground survey points selected in this invention can better represent various possible combinations of "growth-environment" within the crop planting area. Using these ground survey points to validate the remote sensing inversion products of crop growth can more accurately assess the inversion accuracy of the products under different conditions.
[0038] While ensuring representativeness, introducing accessibility analysis based on road networks can significantly reduce the time, manpower, and economic costs of field surveys.
[0039] Each ground survey point output is accompanied by detailed metadata, which not only provides accurate data for remote sensing verification, but is also a complete record of site conditions and growth, providing valuable data support for subsequent analysis of the impact mechanisms of different environments and management measures on crop growth.
[0040] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A method for selecting ground survey points for crop growth remote sensing inversion products, characterized in that, Includes the following steps: Obtain remote sensing inversion products of crop growth in the area to be investigated; Collect and process environmental factor data that affect the spatial differentiation of crop growth, and construct an environmental factor raster dataset; The crop growth remote sensing inversion product is overlaid with the environmental factor raster dataset, and all pixels in the crop planting area are clustered to obtain the candidate regions after clustering. Within each candidate region, patches with an area greater than a preset threshold are identified and retained as retained patches. Edge pixels of the retained patches are removed to obtain core patches. Identify candidate ground survey points for each core patch; For each candidate region, its core patches are sorted, and based on the sorting results, the candidate ground survey points of the top N core patches are selected as the final ground survey points.
2. The method for selecting ground survey points for crop growth remote sensing inversion products according to claim 1, characterized in that, The crop growth remote sensing inversion products include crop leaf area index and / or crop canopy chlorophyll content and / or crop canopy nitrogen content and / or crop canopy carotenoid content.
3. The method for selecting ground survey points for crop growth remote sensing inversion products according to claim 1, characterized in that, The environmental factor raster dataset includes: topographic factors and / or soil factors and / or management factors and / or meteorological factors; The terrain factors include: elevation, slope, aspect, and terrain relief raster data; The soil factors include: raster data of soil type, soil texture, and soil organic matter content; The management factors include: irrigation / dryland zoning raster data; The meteorological factors include: temperature, precipitation, and sunshine duration raster data.
4. The method for selecting ground survey points for crop growth remote sensing inversion products according to claim 1, characterized in that, The specific method for edge pixel removal is as follows: perform one or two rounds of pixel removal where there are no neighbors on one side.
5. The method for selecting ground survey points for crop growth remote sensing inversion products according to claim 1, characterized in that, The method for determining the candidate ground survey points corresponding to each core patch is as follows: Determine the maximum inscribed rectangle, maximum inscribed square, or maximum inner circle for each core patch, and use the center of the maximum inscribed rectangle, maximum inscribed square, or maximum inner circle of the core patch as the candidate ground survey point for the corresponding core patch.
6. The method for selecting ground survey points for crop growth remote sensing inversion products according to claim 5, characterized in that, Determine the maximum inscribed rectangle, maximum inscribed square, or maximum inner circle for each core patch, including: The core patch is converted into a vector polygon, and the boundary of the vector polygon is converted into a graph consisting of nodes and edges, wherein the features of the nodes include their coordinates. The graph is embedded into a multidimensional feature vector, and this multidimensional feature vector is used as the embedding vector of the graph; The vertex coordinates of the largest inscribed rectangle or the largest inscribed square, or the center coordinates and radius of the largest inner circle, are obtained from the embedding vector of the graph.
7. The method for selecting ground survey points for crop growth remote sensing inversion products according to claim 6, characterized in that, The specific method for sorting the core patches is as follows: For each candidate area, it is sorted from largest to smallest based on the area of its core patch, then sorted from largest to smallest based on the area of the largest inscribed rectangle, largest inscribed square, or largest inner circle of the core patch, and finally sorted from smallest to largest based on the minimum distance from the candidate ground survey point corresponding to the core patch to the road network.