For recent decades, especially with the launch of international commercial satellites carrying high-precision sensors, land use dynamic monitoring techniques based on high-resolution remote sensing imagery have undergone a phase of rapid development of key techniques applied in many aspects vastly. This paper is focused on the technological research and procedure of these key techniques, including the process of multi-temporal images, the information classification and feature extraction method, and also the analysis of classified result and driving factors influencing mutual conversion between land classes. Considering the representative meaning as its typicality in China, Yangzhou urban area was regarded as the study zone with multi-temporal images of three periods in the year of 1988, 2002 and 2007. During the procedure of classifying, object-oriented method was applied to e36xtract features from different multi-temporal imagery. Through comparative analysis, it was demonstrated that residential area expanded rapidly especially from 2002 to 2007 and area of the arable and other kinds of land cover reduced considerably. Finally, it is concluded that multiple driving forces affected the land use/change and multivariable linear regression model was used to explore the primary and secondary forces.
Read full abstract