Abstract
Satellite images play an essential role in various applications like geographical information systems, remote sensing, ecology, and oceanography. Synthetic Aperture Radar (SAR) imaging is used to achieve high-resolution images on earth. However, these images are positively affected by unnecessary noises by compression and transmission errors. The noise removal process is a challenging task that it had artefacts and blurring of images. Existing researches and studies proposed various de-noising techniques to improve the accuracy of these images, and that attains specific application. These techniques had not yet attained the high performances due to the inaccurate prediction of objects. The primary aim of this research is to enhance the classification accuracy of different regions like water region, residential area, land region, and forest region from SAR images. From input SAR images, the features are extracted by using proposed hybrid saliency mapping and pyramid histogram of oriented gradients. The most important features are selected by using the Improved Principal Component Analysis (IPCA) technique. Further, the classification of regions is achieved by using a novel forest classifier. The performance of the proposed framework ha analyzed with the measures of accuracy, specificity, sensitivity, precision, recall, and f-score. In the result analysis, the proposed method had achieved 98% of accuracy compared than the state-of-the-art algorithms. From the estimation results, it is concluded that the proposed approach offers better results with increased accuracy for the prediction of different objects in SAR images.