A Methodology For Urban Roads Extraction From High Resolstion Remote Sensing Imagery

The development of remote sensing technology has provided a reliable foundation to obtain rich information on terrain features, but data processing technology is relatively falling behind. There is no doubt that it is hindering the development of3S technology and affecting the update of geographic information database. The road is a very important geographic dataset and it plays an important role in national economy and national defense. Research on the automatic or semi-automatic road extraction has always been regarded as a hot and difficult topic. Although lots of theories and methods have been proposed, there is no one that is widely used in production environment. Since road extraction from high resolution remote sensing images is difficult, in this paper, an automatic urban road extraction method for high-resolution images is proposed. Experimental results of several road images indicate that this method can be used to extract roads.The main contents are as follows:1. Research background and significance of road extraction are introduced. And overview on the basic characteristics and models of the road, then analysis of the difficulties of road extraction from high resolution images. According to this, some research ideas for this paper has been given.2. The basic theory of image segmentation technique is introduced, and the theories and methods of the mean shift image segmentation technology are focused on, and then combine with image fusion technology and Gray-value statistics to achieve the segmentation of road images.3. Defined the shape features of roads which are used to separate the road regions. Then in order to ensure the independence of each road target candidate, a multi-directional morphological filtering algorithm is designed to separate road from the neighboring non-road objects. And experimental result indicate that the superiority of this method.4.Detailed discusses the post-processing of the road, and then a method for broken road line is connected to road network by line fitting and line connection is proposed.5.Proposed a method which combined with mean shift image segmentation, shape indices and the multi-directional morphological filtering algorithm for automatic extraction of urban roads, high-resolution urban image test results show that this method can achieve from the complex environment of the road network, In particular, the higher the extraction accuracy of the linear path. Experimental results of road images indicate that this method can be used to extract roads under complex conditions, especially for the straight roads.