Remote-Sensing Image Classification Using Texture-based Multi-scale Features
National Cheng Kung University, Department of Geomatics, Taiwan
An experiment to study how multi-scale feature improves classification accuracy for remote sensing images. Texture feature is extracted by Semivariogram and SVM is chosen as machine learning method for classification task.