[M-SD41] Geospatial applications for natural resources, environment and agriculture

[MSD41-03] Time series Hyperspectral Analysis for vegetation and tree classification

*Garid Zorigoo¹、Yurino Ishida¹、Yukihiro Takahashi¹、Nobuyasu Naruse² (1.Hokkaido Univ.、2.Shiga University of Medical Science)

キーワード：Remote Sensing, Hyperspectral Imaging, Time-series data, Classification

Using our inexpensive Spectral Scanning device, developed at Hokkaido University, Hyperspectral Images (HSI) of a small forest area were made every day for over 2months. Measurements were taken from September to November in 2021, during which the leaf color-changing period. Since the leaf color changing period is different for each species, we can identify species efficiently. Furthermore, with this type of time series & hyperspectral measurement method, one can have a glimpse of plant phenomenons. Hyperspectral Data analyses are rare compared to multispectral, but Time-series Hyperspectral Data analyses are even more scarce. Developments of data analyses and algorithms in this area are needed since it has 2more dimensions (wavelength and time) than most common multispectral imaging.