The Algorithm Research On The Road Extraction In Images For GIS Application

At present, geographic information systems (GIS) are at a stage of rapid development in theoretical research and application, which has been widely used in resource management, environmental monitoring, transportation, urban planning and other aspect in economic development. Target extraction from images has become the important means of the GIS spatial information updating. As one of the key issues , semi-automatic road feature extraction has attracted wide attention, in the national economic production and military target detection domain this research has a matter of great significance on both theory and praxis.This paper has studied the technique about the road extraction in the transportation image and the remote sensing image. The goal is to identify and extract the road information quickly and accurately from the two types of images .This article is an important part of vectoring large-scale complex images for GIS application. The ultimate goal is to extract road rapidly through interactive, to lower acquisition costs and shorten the processing cycle during vectoring image.Based on researching some theories and methods, this article raises improved algorithms of the road extraction in the transportation image. The main improved method as follows : The image is clustered and segmented based on the color criterion; Extension characteristics complexity algorithm for the characters corrosion solve the characters interference; Based on traditional method of tracking identification, the algorithm add the inside and outside mark, which is to solve some problem, such as the public extraction and repeat, in order to provide better results for the realization of vector images.The method of the road extraction in the remote sensing image is different from the transportation image. This article analysis the characteristic of the road in the remote sensing image and raises improved algorithms of the road extraction in the remote sensing image. The main improved method as follows : We build the road mathematical model and the principles of the road edge point; and establish two obtained criterions of the edge seeds of the road according to the different selecting the point inside or outside of the road manually; The ray-tracing method identify and extract the edge of the road at the core of the seed; On the basis of the traditional repair, this article raises the same-side repair algorithm and the different-side repair algorithm to solve line-break problem effectively and improve the quality of extraction. This paper analyzes the test results of these algorithms, discusses the effectiveness and accuracy, and raises the direction of the feasibility research.