The study on rough set in GIS and remote sensing
Qing-wei Zhu, Dazhi Guo, Hao Zhang · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2005
Recently, spatial data mining and knowledge discovery are developed rapidly. In this paper, first, the developing status of spatial data mining and knowledge discovery are concluded; then deeply studying data mining in GIS and remote sensing are performed; third, Rough sets were executed with TM imagery to carry out data mining and knowledge discovery and finally acquiring feature rules of the relation between soil erosion and three elements: Vegetation cover, Slope and cropland. The paper gives support to land use and land reclaim effective. There are four parts are composed. In the first part some DM background are narrated and some problems hidden, which are very important for us to use, are brought forward. This is to say, much hiding information that is the thing we want most is waste. Then spatial data mining and knowledge discovery (short-hand SDNKD) were appeared in this background. An exhaustive study Dm is beyond the scope of this paper. In this study, we focus on two perspectives 1) integration DM with GIS. 2) Integration DM with RS. In the second part, we describe method of spatial data mining including: Apriori Algorithm, Rough Sets Theory, Inductive Learning, Clustering and so on, and emphasis on Rough Sets. Relationship among RS, GIS and DM are interpreted. Integration GIS with DM and Integration RS with DM are explained by fig1 and fig2 in the third part. An explicit example about soil erosion is depicted to explain the relations through Rough Set and some rules are acquired to give support to land use and land reclaim effective in the fourth part. At last, we conclude our study.