The application of entropy in the urban green information extraction of high resolution remote sensing

Entropy is the important texture feature of an image in spatial domain, also it can be applied in frequency domain to detect the texture feature. QuickBird images of Nanjing, China are considered in this paper. Features of street tree images are extracted using the entropy texture based on GLCM firstly. Then images are transferred into Fourier domain, image features are extracted based on entropy feature by Gabor filter. The advantages and disadvantages of the two methods are analysised, the precision of them are compared and the application future of entropy is discussed.