The Research Of Roads Extraction Based On Wavelet Packet And Mathematical Morphology

With the rapid development of Remote Sensing technology, which we can get more and more aerial and satellite imagery from space. Thus, automated information acquisition from imagery has become a bottleneck, which needs to be overcome. Nowadays, it is important to know the information of road transportation in time for the development of region.The application of the remote sensing technology, which extracts road from the RS image effectively, it is an important subject for research. Its achievements can be applied in the transportation synthetically, and it has guiding function on road traffic administration and planning. It makes planners, manager and policymaker consider more overall, and makes they vision widen. The Information of the road can be got and upgraded effectively in time in ITS system and feasible GIS system, which makes systematic space analytical capacity strengthen greatly, and induces the great development of GIS, ITS system.This article is around drawing the roads from the remote sensing images that have a few wave bands. Through consulting a large number of relevant documents, analyzing and summarizing forerunners' working experience, the author has put forward a kind of method on extracting road relatively suitable from remote sensing images which have different resolutions. With the application of remote sensing, we can extract roads from the RS image effectively. Road extraction from the satellite imagery is studied in this paper and the main works are summarized as below:1. Introducing the background of research and the development trend of road extraction.2. Introducing briefly for wavelet and wavelet packet conception and theory, and getting the road out of RS image on the basis of wavelet packet theory. Having emphasized the advantage of the wavelet packet in extracting the road shape detail.3.Compare the road characteristic on the remote sensing image of resolution ratio of level, set up the description of the roads.4. Introducing briefly basic conception and theory of mathematical morphology, which has incomparable advantages to get out of cockscombstructure around the roads' edge.5. Proposing model make on the basis of wavelet and mathematics morphology to extract road automatically. And compare it with the methods of forerunner.In a word, we can get road from RS image depend on roads' brightness and hue and etc. First, we get a RS image through image processing that get ride of noise, via wavelet package resolution method to raise the spatial resolution of the image. Secondly, rely on the shape of roads; we can eliminate some of interference. Next, we use further method to get pure roads, with morphology we select adapt structure element and use dilate and erode to wrap the saw shape of roads' edge. Through above methods, roads extract from the RS image can realize.