Abstract
Change detection of land cover from satellite images is emphasized in China as a result of rapid land cover changes. This paper discusses a method to calculate land cover changes based on fuzzy settings. In this method, land cover is regarded as fuzzy spatial objects rather than crisp objects. The fuzzy land cover is derived based on a fuzzy classification. The degree of change is then calculated using fuzzy reasoning. In order to minimize any errors in image matching, fuzzy polygons are adopted in the reasoning. By incorporating spectral value differences, the errors in image classification can be reduced considerably. The land cover changes from 1989 to 1998 in a part of Sanya City of China are inferred as an example. The results show that this fuzzy method is able to identify the land cover changes more precisely than the crisp method. Furthermore, the transitional land cover changes can be revealed as a by‐product.
Acknowledgements
This research was funded by the National Key Laboratory of Remote Sensing Engineering of Wuhan University (Project No. 003) and Key Laboratory of Geo‐informatics of State Bureau of Surveying and Mapping (Project No. 010). Also, thanks to the anonymous reviewers for their comments.