Discriminating different landuse types by using multitemporal NDXI in a rice planting area

Research has shown that the use of multitemporal images could obtain better classification over single date images and that seasonal variation of the normalized difference vegetation index (NDVI) could help improve classification accuracy. On consideration of different crop phenology and the seasonal changing characteristics of vegetation, water and bare soil indices from the Landsat Thematic Mapper (TM) images, the present paper uses multitemporal NDXI (NDVI, normalized difference water index (NDWI) and normalized difference soil index (NDSI)) to discriminate different landuse types in a rice planting area. From a comparison of the overall accuracy, Kappa coefficient and producer and user accuracies of different approaches, it is found that the NDXI approach is superior to the traditional classification that uses the original un-transformed images in the discrimination of different landuse types and agricultural land types in the rice planting area. The approach is expected to be more applicable to multitemporal Moderate Resolution Imaging Spectroradiometer (MODIS) images and to be used in discriminating different cropping systems in paddy areas.