面向对象解译方法在遥感影像地物分类中的应用  

Application of Object-Oriented Interpretation to Information Classification on Remote Sensing Image

作　　者：王建芳[1] 包世泰[1] 

机构地区：[1]华南农业大学信息学院

出　　处：《热带地理》2006年第3期234-238,242,共6页Tropical Geography

基　　金：广东省科技攻关项目[2005B30801005];华南农大校长基金[2005K139]

摘　　要：针对高分辨率遥感影像快速高效萃取有用信息这一遥感技术应用的热点问题,探讨了基于面向对象(Ob-ject-oriented)解译方法的遥感影像自动及半自动解译和提取的新思路。文中分析了面向对象解译方法在地物信息分类应用中的优势,并提出了基于此方法的高分辨率遥感影像多尺度信息提取技术流程。具体结合广州市新白云机场开发区IKONOS高分辨率遥感数据进行地物快速提取、自动分类的试验,并对解译效果进行了评估分析,证明此方法在高分辨率遥感影像地物分类中确实高效可行。How to extract the valuable information from high-resolution remote sensing image is a hotspot of application of remote sensing technology. The object-oriented approach presents a powerful automatic and semiautomatic interpretation for remote sensing image. In this paper the advantages of the approach applied to image information extraction and classification are analyzed, and the flowchart of high-resolution remote sensing image information ex- traction is designed. By using this method, as an example, the information from IKONOS image data for the Development Zone of the Baiyun Ai of the classification results is tion remote sensing data class rport in Guangzhou is quickly extracted and classified automatically, and the precision evaluated. The approach has been proved to be feasible and efficient in high-resoluification.

关 键 词：面向对象 影像解译 分类 E-cognition 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]