Inter-Spectral and Intra-Spectral Features for Effective Classification of Remotely Sensed Images

Feature selection and extraction are the crucial steps that help to achieve meaningful classification of remotely sensed images. This paper presents a novel work, which selects a high level set of features from the remotely sensed images, than the conventional methods. The new features introduced in this work are inter-spectral and intra-spectral features. It is observed that these features aid us to differentiate between the characteristic pixels of each class in the image. Different classifiers are fed with different types of features of the images and a comparison of the same is also presented in the paper.