基于灰狼算法优化支持向量回归模型的木材染色配色算法研究

Research on Wood Staining and Color Matching Algorithm Based on Improved Grey Wolf Optimizer-Based SVR

为提高木材染配色的精度和速度,本文对樟子松木材单板进行染色,提取染色单板的光谱反射率作为输入,以支持向量回归模型(SVR)为基础作为预测模型对染料配方进行预测,用灰狼算法对SVR参数进行寻优,并引入非线性收敛因子和新的位置更新策略改进灰狼算法容易陷入局部最优的缺点,以配方相对偏差作为评价指标,与固定参数的SVR模型及其他模型做对比,优化后的模型配方相对偏差为0.177,配色效果相较于固定参数SVR模型的相对偏差0.344、遗传算法优化的SVR模型的相对偏差0.287等具有明显优势,对提高人工速生材的利用具有重要意义。

To improve the accuracy and speed of wood dyeing and matching,the Scotch pine veneer was dyed in this paper.Take hyperspectral data as input,and the dye recipe was predicted based on the Support Vector Regression(SVR)model as the prediction model.The Grey Wolf algorithm was used to optimize the parameters of SVR,and the shortcomings of the Grey Wolf algorithm that was easy to fall into the local optimal solution were improved.The nonlinear convergence factor and a new position update strategy were introduced,and the relative deviation of the dye recipe was used as the Evaluation Index.The optimized model has a relative deviation of 0.177 in formulation and exhibits a significant advantage in color matching performance compared to the fixed-parameter SVR with a relative deviation of 0.344 and the SVR optimized by genetic algorithm with a relative deviation of 0.287.This is of great significance for improving the utilization of artificial fast-growing materials.