CAJ | 학술논문

表面光泽度是天然高分子材料（如木材）及其制品的重要视觉参数之一，实现天然高分子材料及其制品表面光泽度的快速测定对其表面质量的在线控制与评价具有重要意义。为了实现近红外光谱技术对木材表面光泽度的快速测定以及拓宽近红外光谱技术在高分子材料表面质量控制领域的应用，本研究利用近红外光谱技术结合偏最小二乘法对天然高分子材料木材表面光泽度的模型预测值与实测值的相关性进行研究，探讨了近红外漫反射光谱技术快速测定天然高分子材料木材表面光泽度的可行性。结果表明：（1）木材表面光泽度与其近红外漫反射光谱密切相关，说明木材表面近红外光谱特征中包含表面光泽度的信息；（2）通过偏最小二乘法建立木材表面光泽度的近红外光谱预测模型，模型对木材表面光泽度的预测值与实测值的相关系数可达0.90；（3）通过改变采谱光纤与木材样品表面的夹角获得不同的漫反射光谱数据，分别建立不同的木材表面光泽度预测模型发现，采集光谱的光纤与样品表面的角度变化对结果影响不显著，光纤与样品表面夹角为90°时的结果相对较好。
표면광택도시천연고분자재료（여목재）급기제품적중요시각삼수지일，실현천연고분자재료급기제품표면광택도적쾌속측정대기표면질량적재선공제여평개구유중요의의。위료실현근홍외광보기술대목재표면광택도적쾌속측정이급탁관근홍외광보기술재고분자재료표면질량공제영역적응용，본연구이용근홍외광보기술결합편최소이승법대천연고분자재료목재표면광택도적모형예측치여실측치적상관성진행연구，탐토료근홍외만반사광보기술쾌속측정천연고분자재료목재표면광택도적가행성。결과표명：（1）목재표면광택도여기근홍외만반사광보밀절상관，설명목재표면근홍외광보특정중포함표면광택도적신식；（2）통과편최소이승법건립목재표면광택도적근홍외광보예측모형，모형대목재표면광택도적예측치여실측치적상관계수가체0.90；（3）통과개변채보광섬여목재양품표면적협각획득불동적만반사광보수거，분별건립불동적목재표면광택도예측모형발현，채집광보적광섬여양품표면적각도변화대결과영향불현저，광섬여양품표면협각위90°시적결과상대교호。
Surface glossiness is one of the important visual appearance parameters of natural polymer material (wood) and its re-lated products .To realize the fast measurement of natural polymer material surface glossiness is of great significance to the on-line quality control and assessment of its surface .In order to broaden the application of near infrared (NIR) spectroscopy in the field of polymer material surface quality control and realize the feasibility of NIR as a fast measurement of surface glossiness ,the NIR combined with partial least squares (PLS) analysis were used to analyse the correlations of natural polymer material wood surface glossiness between the NIR predicted and lab measured ,and then to investigate the feasibility of NIR to rapidly predict the surface glossiness of natural polymer material wood .The results showed that the wood NIR diffuse reflectance spectroscopy regularly varied with the different wood surface glossiness ,from which we can concluded that the NIR spectrums reflected the information of wood surface glossiness .The correlation coefficients of surface glossiness between the PLS models predicted and lab measured were up to 0.90 .Additionally ,by changing the degree between the fiber and sample surface ,we collected the dif-ferent wood NIR spectrums ,the accuracy of NIR surface glossiness models based on these NIR spectrums had not significantly improved ,and models based on the NIR spectrums collected by the 90 degree between the fiber and sample surface performed better .