Title:	DETECTION OF ROAD IN A HIGH RESOLUTION IMAGE USING A MULTICRITERIA DIRECTIONAL OPERATOR
Author(s):	M. Naouai , A. Hamouda , C.weber
ISBN:	978-972-8924-84-3
Editors:	Yingcai Xiao, Tomaz Amon and Piet Kommers
Year:	2009
Edition:	Single
Keywords:	Object recognition, road extraction, directional operator, anisotropic diffusion, edge detection, valley.
Type:	Full Paper
First Page:	15
Last Page:	23
Language:	English
Cover:
Full Contents:	click to dowload
Paper Abstract:	Availability of very high-resolution satellite images of urban areas is recent. It opens new possibilities for the extraction of the linear structures such as roads. While, the high-resolution allows realistic representation of roads on a map, it generates a significant noise. This paper proposes a new road detection method which makes use only of digital images as source of information. This method allows detecting and determining at the same time the orientation of the roads by exploiting some of their intrinsic properties: the weak variance of the level of the gray, the average brightness and the presence of strong curves in the orthogonal direction. Moreover, an adequate combination of these characteristics, calculated on a relatively significant number of directed directional supports, provides a precise estimate of the local orientation of the linear structures on real images. It is then possible to follow and to extract the roads lines present on an image. Indeed the method uses a strong geometrical constraint in order not to depend on a radiometric profile which is too variable in urban areas.
	Go Back