| Rice is a very important ration in the world,and its cultivation is related to the national food security.Remote sensing technology is an important and effective means to identify and extract rice planting areas.With the continuous increase of data sources,the extraction accuracy is also constantly improved.Jianghan Plain,as an important rice planting area in China,has always been the main grain production area.Therefore,the region has a strong representativeness and typicality,which is very suitable for this study.Sentry No.1 and No.2 data are on the sky one after another,at the same time,they improve the spatial,temporal and spectral resolution of large-scale remote sensing images,and provide new opportunities for remote sensing classification and extraction of rice.In this paper,Jiangling County,Jingzhou City,a typical rice planting area in Jianghan Plain,is evaluated by the method of extracting rice planting area from sentry-1 and sentry-2 satellite data.Four methods of extracting rice planting area are compared and analyzed,and good results are obtained.Among them,the highest accuracy of the four methods is achieved by using the combination of sentinel 1 and sentinel 2 data.This research uses the popular Google Earth engine platform in recent years.Gee platform,which is a very powerful computing resources and massive line data,we can use Google Earth engine A large number of published geographic products and images are called on the platform,and a large number of existing algorithms are provided on the platform to realize online computing.Compared with the traditional local platform,the workload from remote sensing data acquisition to data processing and analysis is greatly reduced and the work efficiency is greatly improved on the GEE platform.After evaluating the classification method,we first learned and studied the key period of rice growth in Jianghan Plain of the study area on the GEE platform,combined with sentinel 2 data and Landsat data series.Then,according to the sampling points of field survey,we calculated the NDVI vegetation index curve of nearly 10 years in Google Earth engine platform.Through the analysis of NDVI curve,it was found that the phenological changes of middle rice in recent 10 years were not significant.In the classification of rice planting area in Jianghan Plain,combined with multi temporal remote sensing image and random forest classifier,the rice planting area in Jianghan Plain from 2010 to 2019 was extracted by remote sensing classification.Finally,on the basis of the classification results in the past ten years,through statistical analysis,combined with relevant statistical data,seven image factors that may affect the distribution of rice planting in Jianghan Plain were selected.Then the spatial autocorrelation analysis of counties and regions in Jianghan Plain shows that there is spatial autocorrelation and spillover effect in the surrounding areas of Jingzhou City.Further analysis and Research on the spatial and temporal changes of rice planting area in Jianghan Plain in the past 10 years and the driving factors of the spatial and temporal changes of rice planting area,such as the resident population,gross product,water area and other 7 factors were carried out. |