Research On Storage And Management Of Agricultural Remote Sensing Big Date Based On Distributed

With the deepening application of remote sensing technology in agriculture,agricultural remote sensing data shows a trend of sea quantification,multi-source,and data fusion.Agricultural remote sensing processing requires the processing of complex data,which poses challenges to the efficient storage and rapid processing of agricultural remote sensing data.This study uses distributed storage technology and combines the characteristics of agricultural remote sensing data storage to design a storage and management model for agricultural remote sensing big data.Parallel processing technology is used to design a parallel visualization processing scheme for remote sensing data,and a prototype system for agricultural remote sensing big data storage and management is implemented.The main research findings and conclusions are as follows:1.Build a distributed agricultural remote sensing big data storage and management model.Based on the storage characteristics of agricultural remote sensing data,an overall architecture of an agricultural remote sensing data storage model was designed based on various distributed storage technologies.Through efficiency analysis of various distributed storage technologies,file classification rules were established.Construct a small file storage strategy and design a data storage process,construct a multi-level index model based on the R-tree spatial index model,and achieve rapid retrieval of remote sensing data through the file index module and cache service module.Establish an overall data read and write process for the storage management model to further ensure efficient storage and fast query of data.The model validation results indicate that this study improves storage efficiency by 15%and query efficiency by 80% compared to a single distributed storage solution.The agricultural remote sensing big data storage and management model based on distributed construction can meet the storage and management needs of agricultural remote sensing data.2.Construct a parallel visualization model and algorithm for agricultural remote sensing data.Adopting parallel computing technology,establishing temporary storage for file segmentation based on distributed storage,implementing an adaptive allocation strategy for computing tasks,using a custom RDD based remote sensing data parallel visualization algorithm,and finally constructing a parallel visualization model for agricultural remote sensing data through task merging and judgment.After testing,the visualization processing performance of this research scheme has been improved by more than 50% compared to traditional visualization schemes in terms of processing speed.Simultaneously,the parallel processing performance is 43% more efficient than the Map Reduce parallel solution.This indicates that it can meet the needs of rapid visualization of multi-scale agricultural remote sensing data.3.Implementation of a prototype system for storing and managing agricultural remote sensing big data.By analyzing mainstream remote sensing satellite data at home and abroad,such as Modis,Landsat,Sentinel,etc.,and selecting remote sensing data products suitable for agricultural production,a scalable programming mode is adopted to achieve one-stop download of various agricultural remote sensing data.Based on the agricultural remote sensing big data storage and management model,a six layer system architecture was designed and an agricultural remote sensing big data storage and management system was constructed.This provides a solution for the storage and processing of massive agricultural remote sensing data,and a reference solution for efficient storage management and rapid visualization of agricultural remote sensing data in the big data era.