CAJ | 학술논문

研究基于近红外光谱技术的木材密度预测。运用基于高斯核变换的非线性偏最小二乘法建立密度预测模型,并且对所建模型的评价参数进行了对比分析。结果表明该方法建立的预测模型能对样品的密度进行有效预测。研究表明样品近红外光谱信息与样品的实际密度值之间不是单纯的线性关系,非线性模型可以更好地表征二者之间的关系。
연구기우근홍외광보기술적목재밀도예측。운용기우고사핵변환적비선성편최소이승법건립밀도예측모형,병차대소건모형적평개삼수진행료대비분석。결과표명해방법건립적예측모형능대양품적밀도진행유효예측。연구표명양품근홍외광보신식여양품적실제밀도치지간불시단순적선성관계,비선성모형가이경호지표정이자지간적관계。
NIR-based wood density prediction was studied in this paper.Predction models were developed by non-linear partial least squares(NPLS)based on Gaussian Kernel method.Results showed that models developed based on NPLS are valid to the samples studied.Results also indicated the relationship between NIR spectra and samples' density is not a simple linear relationship.The relationship is better characterized by a NPLS model.