Information Extraction of Land Cover from TM Imagery Supported by ERDAS IMAGINE Soft Ware

摘要

The purpose of this paper is to explore information extraction of land cover from TM imagery supported by ERDAS IMAGINE soft ware. The data we selected for classifying is half of a Landsat TM scene of the surroudings of Zhalong wetland. Both unsupervised classifier and supervised classifier are used to perform information extraction. The results show that the combined classification method is more effective than either unsupervised classifier or supervised classifier independently. The converted vector layer from the water extraction raster layer is generated as a coverage of Arc/Info soft ware. Further research can be done in Arc/Info soft ware combining with the vector map. The ERDAS IMAGINE soft ware can perform the classification and rater data to vector data easily, practically, accurately and efficiently.