Automated road network Extraction from high resolution Multispectral images

Abstract

A road network is one of the most fundamental data of geospatial newlineinformation. The extraction of road network from aerial and satellite newlineimages is of great importance of urban and transportation planning. Due to newlinethe development of high resolution optical satellite imagery, large volumes of newlinespatial and spectral data are available. In order to take advantages of these newlinedata accurate and efficient information methods are crucial in remote sensing. newlineEven though, manual road extraction has high accuracy still it is not effective newlinein terms of time and cost when the scenes are more complex. The manual newlineroad extraction is even more complex when the scene consists of tall newlinebuildings, large vehicles and trees. So it is more necessary to develop an newlineautomatic road extraction method. A Multistage Framework was used for newlineautomatic extraction of roads from high-resolution multispectral satellite newlineimages. However, it does not detect the road junctions. In this research work, newlinedifferent techniques are introduced to improve the automatic road extraction newline