The Comparative Study of Remote Sensing Image Classification method based on ERDAS

摘要

ERDAS IMAGINE is a remote sensing image processing system developed by the United States.The paper using ERDAS to classified the remote sensing of Land-sat TM image data by supervised classification method and unsupervised classification method,Using the Yushu City remote sensing image of Jilin Province as the trial data,and classified the forest,arable land and water from the remote sensing images,compared the test data of the supervised classification and unsupervised classification method,shows that the supervised classification method can be better to solute the questions "with the spectrum of foreign body" and "synonyms spectrum" than unsupervised classification method,and optimize classification images,improved information extraction accuracy.The application shows the classification result is consistent with the actual situation and it laid the foundation for land to have the rational planning and use.