Optical characteristics of Douglas fir at various densities, grain directions and thicknesses investigated by near-infrared spatially resolved spectroscopy (NIR-SRS)
Optical characteristics of Douglas fir at various densities, grain directions and thicknesses investigated by near-infrared spatially resolved spectroscopy (NIR-SRS)
复制标题
DOI：
10.1515/hf-2017-0213
复制
发表时间：
2018-05
期刊：
影响因子：
2.4
通讯作者：
Te Ma;G. Schajer;T. Inagaki;Z. Pirouz;S. Tsuchikawa
中科院分区：
文献类型：
--
作者：
Te Ma;G. Schajer;T. Inagaki;Z. Pirouz;S. Tsuchikawa
摘要利用近红外空间分辨光谱（NIR-SRS）系统，也称高光谱成像系统，在线扫描模式下，结合卤素光源（101 mm），对道格拉斯冷杉（Pseudotsuga menziesii（Mirb.）Franco]木材获得。在1000-1600 nm的近红外光谱范围内观察到具有不同密度、晶粒方向和厚度的样品。然后，基于稳态扩散理论模型，采用非线性曲线拟合算法，估算出每个测量位置处平行于和垂直于颗粒方向的吸收系数（μa）和约化散射系数（μ s${\mu '_s}$）。在1457 nm处的吸收散射系数和在1002 nm处的两种约化散射系数与X射线密度计测量的木材密度高度相关。3 mm和5 mm样品的相关系数分别为0.953和0.987，而三个光学配置文件相结合。因此，尽管软木具有非均匀的细胞结构，但近红外-拉曼光谱是一种快速、简便的测量软木光学特性的方法。亚表面密度和晶粒方向可以预测与满意的精度的基础上的几个关键波长，而不依赖于多元统计分析。
Abstract A near-infrared spatially resolved spectroscopy (NIR-SRS) system, also called the hyperspectral imaging system, was applied in a line scanning model combined with a concentrated halogen light source (Ø 1 mm), and spatially resolved reflectance images of Douglas fir [Pseudotsuga menziesii (Mirb.) Franco] wood were obtained. The samples with various densities, grain directions and thicknesses were observed in the NIR range of 1000–1600 nm. Then, a nonlinear curve-fitting algorithm was applied based on the steady-state diffusion theory model to estimate the absorption coefficient (μa), and reduced scattering coefficients (μ′s${\mu '_s}$) both parallel and perpendicular to the grain direction at each measurement position. The absorption scattering coefficients at 1457 nm and two kinds of reduced scattering coefficients at 1002 nm were highly correlated with wood densities measured by an X-ray densitometer. The correlation coefficients were 0.953 and 0.987 for 3 mm and 5 mm samples, respectively, while three optical profiles were combined. It can be concluded that NIR-SRS is a fast and simple method for measuring the optical characteristics of softwood, although it has a non-homogeneous cellular structure. Sub-surface density and grain direction could be predicted with satisfactory accuracy based on a few key wavelengths without relying on multivariate statistical analysis.