摘要
为提高木质粉尘火花检测的准确性,利用基于支持向量机(SVM)的物质分类方法检测木质粉尘火花。选取马尾松和杨木粉尘为研究对象,将两种粉尘分组点燃试验,获取火花和灰分的高光谱图像,提取感兴趣区域(region of interest,ROI)内的发光度数据进行预处理。利用感兴趣区域内的数据建立SVM分类模型,分别利用网格搜索法(GS)、遗传算法(GA)以及粒子群算法(PSO)对两类树种的SVM分类模型进行参数优化,并将三种参数优选方法的分类预测准确率进行对比。结果表明,三种优化方法均能够很好地检测两种树种的木粉火花,其中网格搜索法检测准确率明显高于其余两种,更适于木质粉尘火花探测,这为人造板生产过程中能够高效检测木质粉尘火花提供了一定的理论依据。
In order to improve the accuracy of wood dust spark detection, a material classification method based on Support Vector Machine (SVM) was used to detect wood dust sparks. The Pinus massoniana and poplar dust were selected as the research object. The two kinds of dusts were grouped into ignition test to obtain the hyperspectral image of spark and ash, and the luminosity data in the region of interest (ROI) was extracted for pretreatment. The SVM classification model is established by using the data in the region of interest, and the SVM classification model of the two types of trees is optimized by Grid Search method (GS), Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) respectively. The classification prediction accuracy of the parameter optimization method is compared. The results show that the three optimization methods can detect the wood powder sparks of the two tree species well. The detection accuracy of the grid search method is significantly higher than the other two, which is more suitable for wood dust spark detection. It provides a theoretical basis for the efficient detection of wood dust sparks.
作者
谢思熠
徐兆军
那斌
朱南峰
XIE Si-yi;XU Zhao-jun;NA Bin;ZHU Nan-feng(Nanjing Forestry University, Nanjing 210037, China)
出处
《林产工业》
北大核心
2019年第1期20-24,共5页
China Forest Products Industry
基金
"十三五"国家重点研发计划"林业资源培育及高效利用技术创新"中课题"人造板安全生产与污染减控关键技术"资助
课题编号:2016YFD0600703
关键词
木粉尘
火花探测
高光谱
支持向量机
参数优选
分类识别
Wood dust
Spark detection
Hyperspectral
Support Vector Machines
Parameter preference method
Classification recognition
