A Novel Approach for Morphological Attribute Profiles with Partial Reconstruction

Abstract

The classification of satellite images are difficult because of human eyes can see the colour of visible light mostly in three bands (RGB). The main shortcomings of existing system are leakage problem (when regions related to different structures in the image by connecting fake links leads to missing of certain objects), objects expected to disappear at a certain threshold remain present when objects were connected. And then if the small objects connected with large objects some attributes will be mixed together. The proposed work is a framework for providing better classification of satellite images by partial reconstruction along with morphological operations and finally using SVM (support vector machines) classifier.

Keywords: Satellite images, partial reconstruction, morphological operations, SVM (support vector machines), RGB

Cite this Article

Shilpa Rajan, Sincy John. A Novel Approach for Morphological Attribute Profiles with Partial Reconstruction. Journal of Control & Instrumentation. 2019; 10(1): 1–8p.