Classification of coniferous tree species using aerial hyper spectral observation 2

Hyper spectral image have 67 bands from visible to near infrared red. The purpose of this study is to make a high accurate classification map of tree species using those data. We study about an effectiveness of hyper spectral observation that has high resolution of wavelength. We evaluate the effectiveness by comparing multi spectral data of an existing satellite. The study area is Field Science Center (FSC) of Tohoku University in Japan. We produce supervised classification. In the process of this study, we extract the coniferous area to classify easily coniferous tree species. We compare hyper spectral data with multi spectral data. There are visually wrong extractions using multi spectral data. Hyper spectral observation is effective to extract the coniferous area. Using extracted area of coniferous tree, we make a classification map of coniferous tree. In the map, we classify successfully among coniferous tree species. It is important to select suitable bands based on a purpose of classification.