Research On Cartography Generalization Of City Residential Areas Based On GIS

City residential areas data cartography generalization is a particular case in special element cartography generalization research. The realization of cartography generalization needs follow geography data generalization rules to solve the conflictions caused by reduction of expressive space. Hoverer, special distribution law of city residential areas is also an important influence element of cartography generalization. Therefore, special element—oriented cartography generalization is the focus of the GIS research and application.This main work of this dissertation is placed on the research of special element cartography generalization which is sustained by GIS. The author analyses main content and characteristics of the city residential areas data cartography generalization, establishes general projects for different geography elements, adopts Production-Rule method of artificial intelligence to integrate the synthetical knowledge and synthetical operation, and makes an investigation on synthetical cartography oriented to special geography element of city residential areas.The dissertation mainly centers on city residential areas data cartography generalization and research with different levels— theory, method, technique and demonstration. At last, by analysis and evaluation to the experiment data of three scales(1: 50000,1:100000,1: 250000), a satisfactory generalization result is gained. The main research work concludes:1. First of all, the dissertation discusses current characteristics and difficulties of automatic cartography generalization, according to generalization theory and principle, based on the previous research, this paper establishes synthetical cartography system oriented to special geography element of city residential areas.2. Aimed at the characteristic of city residential areas data, the dissertation studies the key technique of city residential area patches' optimum classification selection, apply staged pattern recognition model to the geometric series graded algorithm of city residential area patches selection, improves coarse and simply data caused by geometric series grading. The research makes graded data can clearly andreflect the spatial distribution law and quantitative difference of city residential area, proves the feasibility of staged pattern recognition model.3. The dissertation constructs the significance-regulation based development model, and applies it to the residential area cartography generalization of a certain city, obtains relatively satisfactory results.