Abstract:Crop type identification and mapping products are required for the monitoring of crop growth, risk stress, crop yield and other parameters, as well as the planting structure adjustment, decision analysis of supply and demand, arable land resource security and ecological effect assessment. Remote sensing data have become the most important data source for crop type mapping, and the emerging digital technology also provides a series of new approaches. However, with the advent of smart agriculture era, new demands are placed on crop type mapping with higher spatial and temporal resolution, higher product accuracy and more automated. The object was to provide a review of technology trends, key issues and demand gaps of crop type mapping based on remote sensing. It was concentrated on the main problems and main research work from the three perspectives of small-scale crop type fine identification, large-scale crop type automated mapping and crop type mapping mode change. It was highlighted that crop type mapping products needed more precise, near real-time and higher accuracy on the small scale, mainly using super-high spatialresolution image data, such as one meter or less. Furthermore, it still faced significant challenges to improve crop type mapping accuracy, such as more than 95%, for extracting high accuracy crop phenotypes information to meet application needs. On the large-scale crop type mapping, it needed to be more automated and meet the reliable accuracy, such as around 90%. High spatial and temporal resolution image data were mainly used, such as 2~5d and 10~30m, and also the issues of how to deal with the storage management and analysis were faced when it came to big data, to develop the classification method in a robust manner over the large scale, and to fine a scientific and efficient ground true sample acquisition approach. It was also presented that the pattern of crop type mapping would also shift from confirming monitoring to early prediction and specific crop detection. Moreover, five prospects were proposed from the perspectives of strengthening scientific research and accelerating application, which provided some ideas for the development of remote sensing crop type identification and mapping products that met the different needs of smart agriculture and smart land.
