AUTOMATIC MANMADE OBJECT EXTRACTION FROM LOW RESOLUTION TOPOGRAPHIC IMAGE BASED ON PRIMITIVE OBJECT EXTRACTION

Abstract

This paper presents a generative model based approach to man-made structure detection in 2D (two-dimensional) natural images. The aim of this research paper is to develop new generic framework (model) to extract and recognize manmade objects of topographic image i.e. roads and buildings from low resolution digital image automatically that will assist both experts and non-experts users. It was hypothesized that; “ An improvement in framework to extract and recognize manmade object from low resolution topographic image using primitive objects i.e. rectangle and parallel lines could be achieved through integrating recent advances in edge detection, noise removal, corner detection, ridge detection, segmentation and feature extraction techniques with new edge detection, noise removal, corner detection, ridge detection, segmentation and feature extraction techniquesâ€. These new techniques will require minimal architectural changes or modification to the existing framework.