基于深度学习的四种阔叶材材种辨识研究  

Identification of four hardwood species based on deep learning

作　　者：王涛[1] 杨霄霞 高宜生[2] 葛浙东[1] 刘晓平 周玉成[1] Wang Tao;Yang Xiaoxia;Gao Yisheng;Ge Zhedong;Liu Xiaoping;Zhou Yucheng(School of Information and Electrical Engineering,Shandong Jianzhu University,Jinan,Shandong 250101,China;School of Architecture and Urban Planning,Shandong Jianzhu University)

机构地区：[1]山东建筑大学信息与电气工程学院,山东济南250101 [2]山东建筑大学建筑城规学院

出　　处：《计算机时代》2022年第11期76-80,共5页Computer Era

基　　金：山东省自然科学基金青年基金“基于X射线断层成像技术的树木年轮识别与分析”(ZR2020QC174);国家文物局重点科研基地科研项目“CT成像技术在建筑遗产木构架损伤探测中的应用研究”The Natural Sciences and Engineering Research Council of Canada(NSERC)。

摘　　要：阔叶材材种辨识,需对其染色、制片,在显微镜下观察木材标本结构,这一过程步骤繁琐且耗时长。针对此问题,采用深度学习,实现阔叶材快速、准确检测。基于深度学习方法,分别选择YOLOv3框架以及SSD框架。试验结果表明,基于YOLOv3和SSD框架的四种阔叶材辨识的平均精度值分别为91.57%和91.17%,检测用时分别为1.185s和0.608s。这两种框架都可实现四种阔叶材材种高效、准确辨识,证明深度学习方法可以作为四种阔叶材辨识的技术手段。The identification of hardwood species requires dyeing,making slices,and observing the structure of wood specimens under a microscope,which is a cumbersome and time-consuming process.In response to this problem,deep learning is used to achieve rapid and accurate detection of hardwood.The YOLOv3 framework and the SSD framework are chosen respectively.The experimental results show that the average accuracy of the four types of hardwood identification based on the YOLOv3 and SSD framework is 91.57%and 91.17%,and the detection time is 1.185s and 0.608s,respectively.Both frameworks can realize the efficient and accurate identification of the four types of hardwoods,which proves that the deep learning method can be used as the technical means for the identification of the four types of hardwoods.

关 键 词：阔叶材辨识 CT YOLOv3 SSD 深度学习 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]