Rapid and accurate extraction of crop type,spatial and temporal distribution is of great significance for agricultural structure adjustment and national food security.However,there are few optical remote sensing im-age of cloudy areas,thus crop monitoring is limited.To make up this shortage,spectral signature of winter crops and SAR time series characteristics of summer crops were proposed based on the Sentinel-2 and Sentinel-1 data for high-accuracy crop mapping.The Guanghan County,an important grain-producing region in south-west China,was studied.The object-oriented decision tree classification method was explored for spatial and temporal distribution extraction of crops in study area,and the classification accuracy was verified.The results shows that:(1)the main crops in Guanghan County are grain and oil crops,and the major crop rotation pat-terns are wheat-rice,rape-rice,potato-soybean and potato-corn;(2)the SAR time series characteristics of rice,soybean,corn show clear differences,extracting the types and distribution of winter-summer crops based on the optical-SAR remote sensing images provides a new idea for crops monitoring by remote sensing images in cloudy areas.(3)The overall accuracy and Kappa coefficient of object-oriented method reach 85.49%and 0.81,which can maintain the integrity of large area crops,and avoid salt and pepper noise.
Optical remote-sensing imageRadar remote-sensing imageCloudy areaObject orientedCrop information
NETL
NSTL
万方数据
