|| 要旨トップ | 目次 |||日本生態学会第59回全国大会 (2012年3月，大津) 講演要旨
一般講演（ポスター発表） P3-005A (Poster presentation)
Hyperspectral imaging would be a new strong tool for the non-destructive observation of carbon dynamics in rhizosphere. Spectral signatures in the shortwave-infrared (SWIR) spectral region can provide the information of the absorbance by organic compounds such as lignin and cellulose. In this study, we conducted a rhizobox experiment using young poplar cuttings to clarify the potential of SWIR hyperspectral imaging.
The reflectance images of root, litter and soil of the potted cuttings were captured through a clear glass at 253 spectral bands ranging from 0.9 to 2.4 µm. Then, some samples were collected, and their chemical constitutions were determined in lab for the calibration of estimation model. Concentrations of lignin, cellulose and TC were estimated by using reflectance image data and partial least squares regression model. The estimation error of model ranged from 18.4 to 35.1% of average value. Spatial and temporal variations in the chemical constitution such as increments in lignin with root growth were visualized successfully. Our result indicates the high potential of SWIR imaging for the ecological studies of rhizosphere.