摘要
以樟子松单板为染色基材,采用活性染料对其进行染色,获取3组单色染料染色样本与8组混合染料染色样本。将自适应权值粒子群算法(APSO)与Stearns-Noechel模型结合,对8组混合染料染色样本进行染色配方预测。结果表明:采用APSO预测的配方绝对偏差均小于0.1,约为最小二乘法和传统粒子群算法的1/2;迭代次数约为传统粒子群算法的60%,寻优速度较快。优化模型获得三种珍贵材的R-Y-B染色配方分别为:花梨木—0.192%,0.419%,0.082%;鸡翅木—0.137%,0.284%,0.193%;黑酸枝—0.336%,0.581%,0.204%。采用该配方对樟子松单板进行染色,染色拟合样与标准样间的色差均小于2,且两者光谱反射率间的误差小,染色效果较好。
Pinus sylvestris veneer was used as the staining base material,which was stained with reactive dyes.Three groups of single-dyed samples and eight groups of mixed-dyed samples were obtained.Using the adaptive particle swarm algorithm(APSO)combining with the Stearns-Noechel model,the dyeing formula were predicted for eight groups of mixed-dyed samples.The results showed that the absolute deviation of formula obtained by APSO was less than 0.1,which was about 1/2 of the least square method and the traditional particle swarm algorithm.The optimization speed was efficient,and the number of iterations was about 60%of that of the traditional particle swarm optimization algorithm.The R-Y-B dyeing formulas of the three precious wood species obtained through by the optimized model were 0.192%,0.419%,0.082%for rosewood;0.137%,0.284%,0.193%for wenge wood;and 0.336%,0.581%,0.204%for black rosewood.Using this formula of dyeing camphor pine veneer,the color difference between the dye fitting sample and the control sample were less than 2,and the error between the spectral reflectance of both was small.The performance of dyeing was fine.
作者
管雪梅
李文峰
黄静一
GUAN Xue-mei;LI Wen-feng;HUANG Jing-yi(College of Machinery Electricity,Northeast Forestry University,Harbin 150040,Heilongjiang,China)
出处
《木材科学与技术》
北大核心
2021年第5期18-23,共6页
Chinese Journal of Wood Science and Technology
基金
国家自然科学基金面上项目“基于深度学习的木材染色计算机智能配色模型研究”(32171691)
中央高校项目“基于光谱拟合的木材染色配色模型研究”(257202BF02)
黑龙江省自然科学基金联合引导项目“基于深度学习的东北速生材单板染色光谱拟合配色模型研究”(LH2020C37)。
