Font Size: a A A

Automatic Acquisition Of Urban Roads And Highways From Color Traffic Raster Maps

Posted on:2014-02-07Degree:DoctorType:Dissertation
Country:ChinaCandidate:T HaiFull Text:PDF
GTID:1222330398972851Subject:Systems Engineering
Abstract/Summary:PDF Full Text Request
With the development of geographic information systems for transportation, it is becoming more and more important to rapid establishes a traffic digital map. To generate automatically road vector data in low cost, it is a fast and efficient way to acquire automatically roads from color raster maps because the road vector data is the basic data of a traffic vector map. It is easy to recognize artificially the roads from a color raster map, but it is a challenging task using a computer to acquire automatically the roads from the same color raster map.This dissertation takes color the urban traffic raster maps and the color highway traffic raster maps as the research objects. Based on the comprehensive utilization of the features of the two types of the color traffic raster maps, we focus on the colors automatic acquisition, noise elimination methods and the point-shape symbols automatic acquisition in the color traffic raster maps. For acquiring automatically the roads from the color traffic raster maps, a system is developed and implemented by utilizing these methods proposed in the dissertation.Main contributions of the dissertation include:(1) Utilizing comprehensively the features of backgrounds and roads in a color traffic raster map, the same color block detection method is proposed to automatic acquire the colors of backgrounds, and a run length analysis method is proposed to automatic acquire the colors of urban roads from the color urban traffic raster map, and a noise elimination method and a neighborhood method are proposed to automatic acquire the colors of the highways from the color highway traffic raster map. Then these methods extract accurately the colors of all backgrounds and roads, and solve the manual acquisition colors of the backgrounds and roads.(2) In order to avoid these existed noise elimination methods eliminating lots of noise during the color clustering analysis of a color traffic raster map, some noise elimilation methods are proposed, which include a linear noise elimination mothed, a blocky noise elimination mothed, a noise elimination mothed based the gradient feature and so on. And then these noise elimination methods are integrated rationally to solve the noise elimination questions in the cases that the colors of the roads are known and unknown, which obtains better effectiveness than the existing noise elimination methods. (3) A feature matching method is proposed and implemented for automatic acquiring all county symbols from a color highway raster map, which extract accurately all county symbols from the color highway raster maps with the different colors. Firstly, utilizing the shape and color distribution features of the county symbols in color highway raster map, two groups of feature points are acquired by integrating the legend of county symbol and its Laplace edge image. Secondly, all county symbols are recognized automatically from the color highway raster map, which based on the statistic features of the feature points. Finally, it is showed that the feature matching method is effective and superior by experiments.(4) A computer application system is designed and implemented in rorder to acquire automatically urban roads and highways from a color raster map. First of all, a digital image processing system is realized, which has the basic digital image processing functions. Secondly, for automatic acquisition of urban roads and highways from color traffic raster maps, all methods proposed in this dissertation are implemented and itegrated rationally. Finally, the effectiveness and superiority of all methods proposed in this dissertation are verified by experiments.
Keywords/Search Tags:Color traffic raster map, Digital map, Roads automatic acquisition, Color automatic acquisition, Color clustering, Noise elimination, Feature matching
PDF Full Text Request
Related items