Mapping Spring Canola and Spring Wheat using Radarsat-2 and Landsat-8 Images with Google Earth Engine

Using remote sensing, it is difficult to accurately extract spring canola and wheat planting area with only optical images because both crops have the same growth period and similar spectral characteristics. Besides, optical images are susceptible to cloud contamination. Synthetic aperture radar is sensitive to canopy structure and is hardly influenced by weather; however, it is difficult to distinguish spring wheat and grass due to the similarity of both canopy structures during the major growth cycle. In order to resolve this problem, the present study proposed a method to extract spring canola and wheat by combining Radarsat-2 and Landsat-8 images based on Google Earth Engine. First, spring canola, forest, water and spring wheat and grass (both were regarded as one object) were extracted from Radarsat-2 image. Second, the cropland was extracted from Landsat-8 image. Third, synthetic mapping was carried out to achieve spring canola and wheat extraction. The result demonstrates that spring canola and wheat were successfully extracted with an overall accuracy of 96.04%.