Feasibility Analysis of Automatic Classification on Land Use Change Dynamic Monitoring

Automatic classification has been applied to land use survey for its advantage of objectivity and high speed. The automatic classification results have met to the accuracy requirements in land-use status investigation. However, whether the automatic classification data can be used to monitor land-use changes and how much the error range is. In this context, little research has been done, especially in large region. In this paper, the land use changes are analyzed by two sets of data which are extracted respectively using automatic classification and visual interpretation by human-computer interaction based on two periods of "Beijing 1" micro-satellite imageries. The results show that the land use data obtained from automatic classification can basically meet the needs of land use status investigation, but can not meet the demand of land use change monitoring in this region.