An Ingenious Texture and Shape Feature Extraction in Remote Sensing Images by Means of Multi Kernel Principal Component Analysis With Pyramidal Wavelet Transform and Canny Edge Detection Method

Abstract

In the rapid growth of the digital world, the dealing of remote sensing image is increased day to day in context with the extraction of information. The feature extractions had been an exigent part among the research to classify the remote sensing images for legitimate information reclamation. In such context this paper focus on the extraction of information from remote sensing images by means of classification of spectral classes. Texture and shape is one of the important features in computer vision for many applications. Most of the attention has been focused on texture features with window selection and noise models. This problem can be overcome through Multi Kernel Principal Component analysis with pyramidal wavelet transform and canny edge detection method for extracting feature in high resolute images based on texture and shape. In this paper, proposed Multi Kernel Principal Component analysis utilizes to extract common information and specify common sets of features for further process and reduces dimensionality. Pyramidal wavelet transform is used to extract texture perception for visual interpretation and it decomposes the images into number of descriptors. So texture can be extracted in an image with tree-structured wavelet. Finally, an edge detection technique identifies the boundary regions from the classified remote sensing image, which is taken as shape feature extraction. The performance of this proposed work is measured through peak signal to noise ratio, Execution time, Kappa analysis and structural similarity for a various remote sensing dataset images.