作者: Ian Leiper , Stuart Phinn , Arnold G. Dekker
DOI: 10.1080/01431161.2011.637675
关键词: Substrate (marine biology) 、 Spatial ecology 、 Spectrometer 、 Pixel 、 Reef 、 Benthos 、 Coral reef 、 Endmember 、 Remote sensing 、 Geology
摘要: Studies investigating the spectral reflectance of coral reef benthos and substrates have focused on measurement pure endmembers, where entire field view FOV a spectrometer is single or substrate type. At spatial scales current satellite sensors, heterogeneity reefs even at sub-metre scale means that many individual image pixels will be made up mixture types. If endmember spectra are used as training data for classification, there discrepancy, because mixed signature. This study investigated directly linked to pixel size high resolution imaging systems, by incorporating multiple types into in situ. A total 334 signatures were measured 19 assemblages The analysed separability using first derivative values, discrimination decision tree was designed identify assemblages. Using tree, it possible 15 with mean overall classification accuracy 62.6%.