TEXTURE ANALYSIS FOR RECOGNITION OF SATELLITE CLOUD IMAGES USING ORTHOGONAL POLYNOMIAL MODEL

摘要

The main aim of this work is to classify the satellite cloud images with high accuracy and good recognition rate using texture as a feature. This work is based on Orthogonal Polynomial Transformation (OPT) to extort the low level feature like texture. Texture is a central feature especially for classifying homogeneous images. For achieving high accuracy the multi resolution technique is used with rotation invariant characteristics. Multiresolution is achieved in some existing methods with high complexity rate. In the proposed work complexity level is reduced by means of reordering coefficients in the form of sub bands. Then, the global feature is constructed and the optimized result is achieved for classifying the cloud.