Super Resolution Mapping of Remotely Sensed Images

The fine spatial resolution is the primary condition for better accuracy in the mapping of land patches. Accuracy varies as a function of the spatial resolution for every remotely sensed image. The information is generally extracted from satellite images by classification techniques. In the classification process of satellite imagery mixed pixels are the major source of uncertainty. Many satellite sensors have been launched to improve accuracy but the main limitation is the cost effectiveness and and the atmospheric conditions. The main limitation is the hidden spatial detail in such images. In order to increase the accuracy of land cover estimates, the use of soft classifiers is unavoidable. The un mixing of pixel involves finding fractional proportions of pixel covered by each class component. The fractional maps generated in the proposed method is given as input to the Super Resolution method. The primary interest is to detect the mixed pixels of the satellite images and focuses on their classification employing the texture feature and then improve the resolution by the novel method of maximizing spatial dependency. The feature selection as texture is based on the spatial arrangement of intensities of an image. The texture features such as colour intensity, energy and local binary pattern are adopted to analyze the mixed pixel problem. Ensemble learning technique is used as the proposed approach of classification