Study On Remote Sensing Land Use Change Detection Methods

Being one of primary research directions in remote sensing technology, the technology on remote sensing change detection is useful to improve our ability to recognize all kinds of land use activities and the results caused by those activities. It is important for us to carry on the research on remote sensing change detection method when we try to update our historical land use/land cover graph and take good advantage of the detection results.This dissertation is about to do some research on different remote sensing change detection methods. The main research content is as follows:First of all, the dissertation discusses a new remote sensing change detection which is based on historical land use graph. By combing one satellite image and the vector graph after necessary geometric rectification, we could detect change patches of the satellite image corresponding to patches in the vector graph. Based on these detected changing patches, we make the further analysis to get the exact changing regions. At last, we could update the historical land use graph with the historical land sue graph and the detected change regions.Then, the dissertation discusses an improved algorithm based on post-classification methods. We take advantage of the approach of texture analysis, and make further subdivision on the result of post-classification images. Finally we could reflect cultivated land change in the form of overlaying proportion of pixels.And then, the dissertation discuss one of primary feature on the application of the remote sensing change detection methods, that is, attaining the factual aims by combing the knowledge of other subjects.