Research On Automatic Extraction Of Map Element Based On Splitting And Combining

Getting Geographic datum is a key step in the development of Geographical Information System (GIS), which takes more than 80% of the whole work in GIS development. So, furthest getting Geographic datum automaticly is an important approach to improve the efficiency of developing or updating the GIS.Geographic datum includes grid datum and vector datum, and vectorizing scanned maps is the main approach to get vector datum.Developing an Automatic Vectorization System of Scanned Map for automatic vectorizing of scanned map is an important way to get useful vector datum.This system involves AI,Pattern Recognition,Digital Image Processing and Cartology.Many experts have made a lot of research in this area. Although, a lot of key problems have been left unsolved and the current vectorization systems are still in the experimental phase.Based on the conclusion of existing problems in this area home and abroad, this article has made some research and analysis on the problems of developing scanned map automatic vectorizing system. The author has made some work on the experience of Artificial Vectorizing in the following aspects:1) Common binary image noise filtering algorithms have been compared. According to the real request for developing an Automatic Vectorization System of Scanned Map, an noise filtering algorithm based on model matching has been designed;2) An algorithm for automatic extraction of map element based on splitting and combining was proposed and map element classes based on different thinning results were also designed;3) According to the request of map splitting, Proper complementarity for Zhang Algorithm for Thinning Digital Patterns has been made,and the complemented algorithm for thinning is able to satisfy the application;4) Software for the above contents has been implemented and the...