Monitoring On Urban Spatial Expansion Based On Remote Sensing Images

With the fast development of the society and economy ,cities have became the principal part of human's modern life,and it's developping and changing with each passing day. Remote sensing has the character of shotcut and convenient, and it's propitious to supervise the expand of cities. Monitoring on urban spatial expansion based on Remote Sensing (RS) images, it is useful to hold the development and change of city, and it is important to country resource manage, entironment supervising and urban planning.The method of urban expand monitoring with remote sensing is in the phase of manual interpreting assisted by computer classification, so there are problems of lower efficiency, higher artificial factor, and unsuitable to standardization and automatization. After analysesing the development states and current developing of the urban expand monitoring with RS image, this paper analyses on the multi-spectrum distributing and region features of urban land-use class, presents a recognition methods based on Region Multi-Center (RMC) for urban land-used, studies a change detection methods of RS image for urban spatial expansion, and its' application on the Chang-sha, Zhu-zhou and Xiangtang Cities (CZTC) expand change dection. Some valuable achievevements are obtained. The contents of this thesis are organied as follow.Chapter 2 summarizes the basic pattern of urban spatial expansion, and its' feature on the resource satellite image, firstly. forms multi-center of land use/cover class by clustered with max-min algorithmic ,and studies the distributions of the distance from the feature to the classificatory recognition. Analyses the distance distributing of urban land-used pixiel from its average centeral, calculates the urban land-use in-class's and between-class's distances from the Land Cover/ Use (LCU)recognition with multi-center and single clustering center, discusses the mechanism of the classification methods such as C-mean, maximum likelihood and fuzzy neural network in the end. A whole analysis method of multi-spectrum feature for urban land-used is presented.Chapter 3 Aiming at the problem that the multi-spectrum feature of land use/cover class is not always normal distribution, because the class is made of multiple covered species, analyses the feature of region multi-center of urban land-use form the other LCU class, presents a recognition method of urban land-use with multi-spectrum remote sensing image based on RMC. The recognition feature of urban land-use has more difference than the feature of other class in the RMC, and RMC transacts preferably the combination of region and spectrum feature, so the method improves effectively the recognition precision of urban land-use.Chapter 4 analyses the development states and current developing of the change detection technique of RS image, With a view to the multiformity and region character of urban expand, and the problem of the error piled up by the approach of classification after comparison, presents a change detection method of comparing before classification for city expand. This method has less classification numbers than the method of comparing after classification, deals with the problem of the classification accumulated. A perfect achievevement is obtained by the application of RMC method to class the compared RS image for city expand change detection.Chapter 5 introdces the data character and disposal method of the results of the city change detection firsthly, sets up the RS Monitoring System of CZTC based on combination RS and GIS, with the system analyses the urban expansion character, current developing, influence factors and expanded problems of the CZTC, and gives the advice of the developing stratagem and city planning of the CZTC. The application indicate that the methods what the paper brings forward reaches basicly applied level.Chapter 6 summarizes the general work of this paper, and points out the further research.The approach of monitoring on urban spatial expansion based on remote sensing images provides a new technical means to study urban expansion and development.