An Approach for Road Extraction from High Resolution Imagery Based on Radon-Like Features and Mathematical Morphology S. Natarajan 1 , P. N. Anil 2 1 Dept. of Information Science, PES Institute of Technology, Bangalore, Karnataka, INDIA 2 Dept. of Mathematics, Global Academy of Technology, Bangalore, Karnataka, INDIA Abstract - Road Extraction from high resolution imagery is of fundamental importance in the context of spatial data capturing and updating for GIS applications. In this paper we present a new approach for extracting roads from high resolution imagery using Radon-like features and mathematical morphology. In this approach first image is preprocessed using bilateral filter and then Radon-like features are used to enhance the road edges. Finally roads are extracted using morphological operations. The method is successfully implemented using high resolution satellite and areal imagery. Keywords: Bilateral filtering. Radon-like features, Mathematical Morphology, Road extraction 1 Introduction Road extraction from remotely sensed data is a challenging issue in the field of photogrammetry and digital image processing. Extensive research has been done on road extraction from aerial and satellite imagery. The methods for road extraction can be mainly divided in to two types, semi-automatic method and fully automatic method. Semi-automatic feature extraction is an interactive process between an operator and computer algorithms. In such methods an operator selects initial point(s) and a direction for road tracking algorithm. Reference [1] proposed a semi automatic road network extraction method based on dynamic programming. Reference [2] proposed a semi-automatic method for urban road extraction based on level set method by using data fusion of multispectral and microwave radar images. In their paper the fast marching method of level set (LS-FM) is used as a tool to fuse different image features for road extraction. Reference [3] proposed a semi automatic road extraction algorithm using template matching. In this method user needs to input an initial seed point to extract a road. Then the orientation of road seed is calculated automatically. They pointed out the method may not work on the roads casted by shadows. Reference [4] proposed a semi automatic road extraction method using multi spectral high resolution satellite images. Firstly, ‘road mask’ was created by multi spectral data classification. Chains of edge pixels were tracked based on local edge direction and straight lines were obtained. Template matching was then used to determine the direction of the line and to obtain next road node. A result in urban area was good for major roads whereas small roads were missed, as road boundaries were unclear due to the objects surrounding the roads. Reference [5] proposed a semi- automatic method for extracting roads from high resolution (1 meter) pan sharpened multispectral IKONOS imagery. In their method an operator provides an initial seed point on the road of interest, then the region is extracted using level set method. The new framework for semi-automatic feature extraction was proposed in [6] and applied to highway extraction and vehicle detection from multiple frame aerial photographs. The basis of new framework proposed in their work is a geometric deformable model. Reference [7] proposed a method to extract the street network as surface elements from topologically correct graph using multi-resolution snakes. Automatic feature extraction has been always an interesting subject for researchers. In the recent years the automated extraction of roads has drawn considerable attention due to the need for the efficient acquisition and updating of road data for geodatabases. Reference [8] presents an automatic approach for the extraction of roads from high resolution multispectral satellite imagery. Lines are extracted in all image channels and employed as initial road hypothesis as well as for the generation of training areas. The goal is to calculate membership value for the road class pixel. The assessment of road hypotheses is done based on geometrical and spectral properties by finding fuzzy values for the parameters length, average width and road energy. Road network is generated using weighted graph and detour factor to close small and large gaps respectively. Reference [9] proposed an automated image processing technique to extract control points for