OBJECT-BASED LAND COVER CLASSIFICATION IN HIGH SPATIAL RESOLUTION REMOTE SENSING IMAGERY OF MOUNTAIN AREA, A CASE STUDY IN MIYUN RESERVOIR AREA

摘要

Based on high resolution remote sensing imagery and in combination with multi-temporal imagery, this article classified land cover of Miyun reservoir area using the object-oriented classification method. The results showed that the deciduous broad-leaved shrub, deciduous broad-leaved forest, dry land, and grass accountedfor 92% of the total area. The study proved that multi-temporal satellite images are critical in grouping those objects with different spectrum together, like the cultivated land. Moreover, plenty of ancillary data is good to distinguish the different objects with same spectrum.