CAJ | 학술논문

面向对象的遥感影像分类方法克服了传统基于像元分类方法的弊端，将对象光谱、空间纹理等特征一并加入分类依据中，有效避免了“同谱异物”或“异物同谱”的问题，适合于高分辨率的遥感影像分类。以武汉市某街区公共遥感影像为例，采用上述方法，结合支持向量机分类方法进行地物分类识别，结果显示，分类总体精度达到了89．9913％，取得了良好的分类效果。
면향대상적요감영상분류방법극복료전통기우상원분류방법적폐단，장대상광보、공간문리등특정일병가입분류의거중，유효피면료“동보이물”혹“이물동보”적문제，괄합우고분변솔적요감영상분류。이무한시모가구공공요감영상위례，채용상술방법，결합지지향량궤분류방법진행지물분류식별，결과현시，분류총체정도체도료89．9913％，취득료량호적분류효과。
Object oriented remote sensing image classification method overcomes the shortcomings of traditional pixel-based classification method .It combines the spectral ,spatial elements texture features of the images ,can avoid the "different objects share the same spectrum"or"the same objects differ in spectrum"effectively .It is suitable for the high resolution remote sensing image classification .Using a public image in Wuhan City as an example ,with the above method ,we com-bined with support vector machine classification method detect terrain classification recognition .Results show that the o-verall classification accuracy is 89 .9913% ,the classification has achieved good results .