A Novel Algorithm to Generate the Cover Map of a Remotely Sensed Image for GIS Applications

In this paper, we present a procedure to approximate the cover map generated from a remotely sensed image. The procedure can detect the nodes and branches from a segmented description of an image resembling a network. In the first of a two part algorithm, the nodes and branches of the mesh are detected. Once this has been completed, all those branches encapsulating a segment are used to generate the cover map. Dominantpoint detection algorithms have been used to detect the breakpoints on the individual branches. This allows the entire scene to be represented by a set of polygons.