Object-Oriented Information Extraction Using HJ-1 Remote Sensing: The Case on Changbai Mountain, Northeast China

Land cover information extraction is the basis of monitoring of land cover change and the environment. In this paper, the Changbai Mountain region was selected as the study area. The object-oriented information extraction technique, the multi-temporal HJ-1A images, and other auxiliary data were used to extract the main land use/land cover types of study area. The classification accuracy were evaluated through ground sampling data. Comparing with the pixel-based classification method, the results of this study has relatively higher accuracy. Especially, the multi-temporal images of HJ-1A play an important role in the extraction of forest types. DEM, vegetation index and other index obtained from different band calculation were also proved to be the more valuable auxiliary information for improving the classification accuracy in mountain area.