Classification of Bamboo and Evergreen Japanese Bay Trees From Coastal Sanriku Forests of Northern Japan

There are concerns about the lack of forest management in the tsunami-damaged coastal areas affected by the 2011 Tohoku earthquake. In this study, appropriate band combinations and data acquisition times for Sentinel-2 data were examined to extract bamboo and evergreen vegetation from this area. The data obtained from the WorldView-3 and Preiades-1A high resolution optical sensors were also analyzed. The analysis revealed that the classification using short wave infrared (SWIR) demonstrated higher accuracy than the red-edge 3 or near-infrared (NIR) narrow band in Sentinel-2 data analysis. Additionally, the data obtained before the development of new leaves proved useful for extracting bamboo and evergreen forests.