面向对象的高分辨率遥感影像信息提取技术

Feature Extraction of Object-oriented High-resolution Remote Sensing Image

介绍了单张高空间分辨率的航空遥感数据在没有其他专题图辅助的情况下进行面向对象的信息提取的方法。以山西省平顺县为研究区,ADS40彩色航空影像为数据源,对不同的地物采取不同尺度的多尺度分割,形成具有"同质性"的对象,然后分别采取不同的信息提取策略,得到研究区信息分类成果,且有效避免了"同谱异物"、"同物异谱"及"椒盐现象",对高分辨率遥感影像信息提取技术研究具有一定的实践指导意义。

This paper focused on object-oriented information extraction technology research with a high spatial resolution airborne remote sensing data that was absence of other thematic map aided information.Taking Pingshun County of Shanxi Province as a study area,and ADS40 color aerial images as data sources,different objects were segmented to homogeneity of different scales.Then it adapt different information extraction strategy to get the classification results,and effectively avoided the'different bodies with the same spectrum','content with different spectrums'and'pepper phenomenon',which has a certain practice meaning in information extraction from high-resolution remote sensing image.