Land Cover Classification from Remotely Sensed Imagery Using Computational Intelligence with Application to the Heihe River Basin

摘要：

To develop and apply framework for unsupervised and supervised land cover classification from multi-band satellite imagery for the Heihe River Basin. The supervised classification mainly uses probabilistic methods, such as Maximum Likelihood Classifier and Fisher Linear Likelihood and the Minimum Euclidean Distance method. The unsupervised classification used is based on two parts, namely: Feature Extractor and Self-Organising Feature Map (SOFM) classifier. According to the performances of the supervised classification methods, the Maximum Likelihood performs better than the Fisher Linear Likelihood and the Minimum Euclidean Distance method. Notably, the supervised classification algorithms demand larger land cover information about the ground truth in comparison with the SOFM classifier for the automated remotely sensed image analysis.

关键词： land cover classification supervised classification Self - Organising Feature Map remotely sensed imagery Heihe River Basin

作者: Wang Chunqing Zhang Yong Yang Jinfang

作者单位: Hydrology Bureau of YRCC, Zhengzhou, 450004, China UNESCO - IHE Institute for Water Education, 2611 Hydrology Bureau of YRCC, Zhengzhou, 450004, China

会议类型: 国际会议

会议名称: The 4th International Yellow River Forum on Ecological Civilization and River Ethics(第四届黄河国际论坛 2009 IYRF)

会议地点: 郑州

会议语种:英文

页码: 1502-1510

在线出版日期: 2009-10-20（万方平台首次上网日期，不代表论文的发表时间）