Feature extraction and object detection in remote sensing images using bow and fast cnn based on unsupervised classification models

Abstract

Remote sensing images have been successfully exploited to deal with several applications including Land cover control, Forestry management, Urban planning and development, Climate change analysis, Soil Mapping, Ocean resources and Natural disaster recovery. More than a dozen earth observation satellites are providing very high resolution remote sensing images with a great number of spatial data, but only few fields are utilizing these effectively, because it is very difficult to process thousands of image data and extract useful information from them. Feature Extraction and Object Detection being the fundamental but challenging problem in the field of remote sensing image analysis plays an important role and is receiving significant attention in recent years. In this thesis, new methodologies of feature extraction and object detection are proposed and utilized to improve the performance of remote sensing images. newline