Mainly introduces the integrated implementation of the basic principle and the algorithm of object image analysis and minimum distance classification method,and uses the Worldview Ⅱ satellite images of Zhongwei for classification of land,and the classification results compared with the traditional minimum distance classification method based on pixel,visual explanation translation and quantitative evaluation results all show that the indexes of information extraction method based on object is more superior,the overall accuracy from 0.85 to 0.87,kappa coefficient by 0.81 increased to 0.84.Therefore,for high resolution remote sensing image,integrated minimum distance classifier,extraction method based on object is more superior to the method based on pixel,classification results with higher precision.%主要介绍了集成基于对象的影像分析与最小距离分类方法的原理,采用中卫市WorldViewⅡ影像进行土地覆盖分类研究,并将分类结果与传统的基于像元的最小距离分类结果进行对比.目视解译与定量评价均表明:基于对象方法的各项指标更优越,总体精度由0.85提高到0.87,Kappa系数由0.81提高到0.84.因此,对于高分辨率遥感影像,集成最小距离分类器,基于对象的信息提取方法要优于基于像元方法,分类结果精度更高.
展开▼
