Monitoring Non-grain Use of Croplands on the Jianghan Plain Based on Sentinel-2 Vegetation Phenology

Objectives:  Monitoring the "non-grain" use of cropland using remote sensing technology is of great significance for national food security and rural revitalization.   Methods:  Through analyzing the phenology of different crop patterns by using time series Sentinel-2 remote sensing images, several key phenological stages were selected to summarize the phenological characteristics of each stage, and the phenology similarity of the same crop pattern and the phenology difference of different crop pattern were obtained. A model was built to extract the nongrain use of cropland on the Jianghan Plain, including vegetables planting, abandoned land or trees planting and aquaculture pond, etc.   Results:  The method realized the mapping of non-grain use of cropland on the Jianghan Plain successfully by using hierarchical classification method according to the rule of "from simple to complex". The overall accuracy of the results is 92.69%, and the Kappa coefficient is 0.89.   Conclusions:  The experiments demonstrated that the method based on time-series Sentinel-2 remote sensing images and crop phenology can be used to monitor non-grain use of cropland at regional scale. The proposed method can provide effective technology for monitoring the non-grain use of cropland, and provide basic data for agricultural departments to formulate agricultural policies.