Road tracking from high resolution satellite images using a new set of profiles and Bayesian filtering

In This paper, a new semi-automatic method for road extraction in urban or non-urban areas is presented to produce a geographical map and updating it. We have tried to develop a new method for road tracking based on road features in high resolution images. To do so, inspiring from profile matching that is usually used for road tracking; we have defined a new set of profiles. This set was able to detect roads in different places such as urban or semi-urban areas. These profiles should have some distinct characteristics to follow the correct path which are thoroughly investigated here. Our proposed method is considered as a semi-automatic method, which the start point is fed to system by the human operator. Moreover, Bayesian filters such as extended Kalman filter and particle filter is added to system to increase its reliability.