Extraction of buildings from remotely sensed images

Abstract

Automatic Building extraction plays a significant role to become newlineestranged the buildings from the background regions in high resolution newlineremote sensing images. It is a binary classification process. Building newlineinformation is necessary for several applications such as mapping, urban newlineplanning, population assessment, disaster monitoring, and infrastructure newlinemanagement. The detection of buildings and other objects became possible newlinewith the development of very high-resolution remote sensing pictures. Highresolution newlineremote sensing photographs provide detailed spectral and textural newlineinformation, as well as high temporal and spatial resolution. Important newlinefeatures are required to analyze the high resolution satellite images instead of newlinepixel information. The prominence of this research is to design and newlineincorporate Bi-Channel and Bi-Spatial (B2-CS) LinkNet model for performing newlinepixel-wise labeling of aerial images that improve the segmentation accuracy newlineof building extraction system. newline