基于优化相关向量机的人造板厚度在线检测

Online Detection of Wood-based Panel Thickness via Optimized Relevance Vector Machines

针对人造板厚度检测系统检测精度不高的问题,提出一种基于改进麻雀搜索算法(Improved Sparrow Search Algorithm,ISSA)优化相关向量机(Relevance Vector Machine,RVM)的人造板厚度检测方法,以提高人造板厚度检测系统的检测精度。从两个角度对传统麻雀搜索算法进行改进:首先在初始种群位置中引入精英混沌反向学习机制,使算法的初始种群分布更加合理,提高了初始解的质量;然后通过引入一种变尺度混沌变异算子,对停滞的全局最优解进行变异,以增强算法的抗停滞的能力,在此基础上通过改进后的算法优化相关向量机的核函数参数,最后以中密度纤维板(Medium Density Fiberboard,MDF)为例开展了在线检测试验,获取试验数据并进行对比分析。结果表明:所提方法能够有效减少检测误差,提高测量精度。

To deal with the issue of low detection accuracy of wood-based panel thickness detection system,a wood-based panel thickness detection method relying on improved sparrow search algorithm(ISSA)and optimized correlation vector machine(RVM)is proposed to improve the detection accuracy of wood-based panel thickness detection system.By improving the sparrow search algorithm from two perspectives and introducing the elite chaotic reverse learning mechanism into the initial population position,the initial population distribution of the algorithm is more reasonable and the quality of the initial solution is improved;Then,a variable scale chaotic mutation operator is introduced to mutate the stagnant global optimal solution to enhance the anti-stagnation ability of the algorithm.The improved algorithm is used to optimize the kernel function parameters of the correlation vector machine.Finally,the on-line detection experiment is carried out with medium density fiberboard(MDF),and the experimental data are obtained and compared.The experimental results show that the proposed method can effectively reduce the measurement error and improve the measurement accuracy.