Feature-based Land Use/Land Cover Classification of Google Earth Imagery

In this paper, we presented a novel method to classify land use/land cover objects of earth surface using google earth imagery. The Gray Level Co-occurrence Matrix (GLCM) is used to extract second order statistical features. Opening, closing, and reconstruction operators are used to extract morphological features of objects. The extracted texture features and morphological features are fused to improve the classification accuracy. The spectral reflectance of the endmember classes is calculated using Linear Spectral Mixture Model (LSMM). Multiple thresholds are generated for each class. The generated spectral properties and thresholding are used to classify features set. The classification accuracy of the resulted thematic map of the specified geographical area is compared with the generic KNN method.