Extraction of trees and grasses in urban green land based on texture character of high resolution remote sensing image

Application high resolution remote sensing image, based on multiscale segmentation and object-oriented image analysis method, and vegetation index was computed, green land was extracted. In order to distinguish different types of vegetation, applying texture character classification algorithm of grey co-occurrence matrix, trees and grasses were extracted respectively. By accurate evaluation, classification had higher accuracy, the result revealed that the way could derive trees and grasses exactly.