摘要
遥感图像的监督分类常用方法目前可以分为:平行六面体法,马氏距离法,最大似然法,神经网络法以及支持向量机法等。本文将就以上所述的五种常用的监督分类方法在ENVI中分别对汶川县威州镇同一Landsat8 OLI数据进行土地覆盖与利用状况分类.比较各种方法的分类精度,并对之所产生的差异的原因进行浅析,进而对实际的生产以及应用做出借鉴。
Remote sensing image supervised classification methods commonly used nowadays are: Parallelepiped, Mahalanobis Distance, Maximum Likelihood, Artificial Neural Net and Support Vector Machine Classification etc. This essay compares five commonly used supervised classification methods introduced above using the same Landsa8 OLI data of Weizhou, Wenchuan to solve the land cover & use circumstance. Comparing the accuracy among the five methods, and briefly analyze the reasons that causes the differences of different methods. The Conclusions may be used as references for the practical production and application.
引文