The Research Of Road Extraction From High-resolution Remote Sensing Takes Into Account Of Geometric Features And Mathematical Morphology

With the rapidly development of remote sensing techniques and computer techniques, the high-resolution remote sensing can provide people with high quality and rich source of data. Road network is an important foundation for the basic geographic information elements. The high-efficiency and high-precision extraction of road has significant value in the cartography and updating of data. High-resolution remote sensing has become an important means of access to the road information. Road information in remote sensing shows the following characteristics:the geometric characteristics of the same width and long length; the radiation characteristics of gray slow changes in the local area; the spectral characteristics of strong robustness; the topological features of the network connection between the roads; the road functional characteristics can be realized in reality and the contextual features of roads associating with surrounding area. Therefore, the extraction of road network is an important aspect of remote sensing object extraction from hyper spectral remote sensing images.Mathematical morphology is a branch of applied mathematics. It is an image processing analysis tool which is studying the spatial structure of the target image by detecting morphological characteristics of the target. The basic idea of mathematical morphology is to use certain forms of structural elements to measure and extract the corresponding shape of the images in order to achieve the purpose of analysis and recognition of the right image. Mathematical morphology has been widely applied in the field of computer vision, image processing and analysis, pattern recognition, surveying and mapping and so on. Based on the theory of mathematical morphology and geometric characteristics of the road, in this paper it is focused on the extraction of the road network from the high-resolution remote sensing images. The main works of article are presented as follows:1. Introducing the significance of road information extraction, and the application of the mathematical morphology in image processing and road extraction in the domestic and foreign research profile. 2. Introducing the theory and basic operations of mathematical morphology. The main content among this part is the theory and algorithms of binary morphology and gray scale morphology, and the algebraic properties of these operators are described.3. The difficulties and particularity of road extraction in high-resolution remote sensing image are analyzed. Then using morphological opening and closing operation, Top-Hat transform, threshold segmentation, form refinement to achieve accurate extraction of the road information in analog image.4. Take full account of the geometric characteristics of the road on the basis of in-depth research and analysis on the road characteristics of remote sensing image, using the mathematical morphology algorithms that achieve the accurate road extraction in the analog image to the road extraction in the experimental image. Using the morphological opening algorithm to eliminate the noise, using the morphological closing algorithm to fill the loopholes noise during the road extraction process, using threshold method to segment image, at last, using the form refinement to extract the skeleton of the road, then we can get a complete road network extraction result.The study showed that using mathematical morphology to extract road information in high resolution remote sensing images based on taking full account of road geometric characteristics of remote sensing images can get better extraction and high geometric accuracy.