Study On Land Use/Land Cover Change By Multi-spectral And SAR Images

With the rapid expansion of population and the development of science and technology, the scope and intensity of land use has been unprecedented,which has caused in a series of global environmental issues. Land use /land cover change will be the dominant factor in the global environmental change. It becomes more and more important to the global ecosystems, human health and human wealth. Monitoring land use / land cover change is very important for the sustainable use of land resources and for harmonizing ecological and economic goals as well as environment..Multi-sensors have been on orbit, which provide effective ways for monitoring land use / land cover change. In the thesis, a case study in the area of Qingdao is carried out. The land use and land cover are classified using images obtained from satellite multi-spectral imagers and Synthetic Aperture Radars (SAR). The results are given below.The results using unsupervised classification and supervised classification algorithms including k-means,isodata,spectral angle mapper,minimum distance,neural net and maximum likelihood method are compared. Best performance is using the maximum likelihood method.The maximum likelihood method is selected as the classifier. The usage of multi-spectral images and combination of multi-spectral and SAR images are performed respectively in the classifications. The results indicates that the accuracy of the classification is better combined multi-spectral and SAR images.The land use/land cover changes are obtained by Post-classification comparison and composite of SAR images. The change of different uses is obvious by using Post-classification comparison, but the accuracy is affected by both two classification accuracy. The composite of SAR images are more sensitive to the large area of land-use change, but change of different uses is unobvious, need manual interpretation and field verification.