As human activities expanding and the process of urbanization in the past decades

urban land use changes very quickly at different scales in China.Extensive studies have been carried out to extract information of land use changes from remote sensing data.Conventional remote sensing change detection methods such as direct comparison and post-classification comparison are performed at pixel level.However

these methods have been proved to be less effective in quantitatively detecting subtle changes within one land use class than detecting land use transitions

i.e.qualitative changes occurred between different land use classes.To enable the capability of detecting quantitative changes in urban land use

a change detection method is proposed based on impervious surface mapping with multi-resolution remotely sensed data.Urban development leads to the increase of impervious surfaces in urban areas

and the impervious surface has been recognized as an important urban land cover type and one of the key factors in the land

hydrological

climatic

ecological and environmental studies.In this paper

the classification and regression tree（CART） algorithm is used with both high-resolution（QuickBird） and medium-resolution（Landsat5 TM） remote sensing data to establish prediction models of impervious surface percentage（ISP）.Based on bi-temporal results of ISP prediction

urban land use changes from 2002 to 2006 are detected in Tai’an city of Shandong province.Furthermore

preliminary analysis for these urban land use changes is carried out.The experimental results demonstrated the feasibility and effectiveness of this change detection method which can be used as a supplement to conventional change detection methods.