Extraction of Roads from Remote Sensing Imagery Using Pixel based and Region based Methods

Abstract

The development of high resolution imagery from satellites and the number of newlineavailable aerial images are exploded in recent years. The need of analyzing these data has newlinegrown immensely. Unfortunately, the technology for analyzing all of these images has not newlinebeen able to keep up-to-date and much of the work of analyzing the images is still newlineconducted manually by humans which are expensive, time consuming and error prone. newlineBecause of these, there is a high demand for fast and reliable methods that are able to newlineanalyze the images automatically. An area where the above issue is prominent is in road newlinemapping. Road maps are essential to our day-to-day life; so they become an important tool newlinein many areas such as in urban planning and automotive navigation. Since the global road newlinenetwork is ever changing, they have to be kept up-to-date. The development of a new newlinemethod for extracting road networks which is both automatic and reliable is therefore newlinenecessary. In order to discern the perfect road extraction mechanism, it is essential that the newlinedetails of road features, their singularity, and the context of their application is understood newlineas a base. newlineRoad extraction from Remote Sensing images stands as an essential node in the newlinerecent research for the development of rudimentary layers in innumerable fields. From newlineGeographical Information Systems to Unmanned Aerial Vehicles, road maps pave the newlinefoundation for data accumulation. This significant process is a result of number of newlinemechanisms devised over the years through iterative experiments and researches newline