作者: Shanshan Li , Xiuping Jia , Bing Zhang
DOI: 10.1109/IGARSS.2013.6723581
关键词:
摘要: The paper presents a supervised classification method based on superpixels and Markov random field (MRF). Hyperspectral image is over-segmented into that are as basic unit of instead operating at the pixel level. Adaptive weight coefficient introduced to determine contextual relationship between superpixels. Support vector machines implemented for better estimation spectral contribution this approach. An experiment real hyperspectral reveals efficient performance.