木材科学与技术2023，Vol.37Issue(5)：61-68,8.DOI:10.12326/j.2096-9694.2023032
一种基于YOLOv5和轮廓约束的木材空洞缺陷应力波层析成像算法
A Stress Wave Tomography Algorithm for Detecting Holes in Wood using YOLOv5 and Constrained Contour
摘要
Abstract
In this study,we proposed a stress wave tomography algorithm based on YOLOv5 and constrained contour for detecting holes in wood.This method ensures the quality of images with reduced number of sensors.Firstly,virtual defect images generated by simulation software were used as samples for training.Next the position of the defect was accurately detected through the YOLOv5 model.Then,the proposed constrained contour algorithm was used to perform global and local constraints on the detection results,to reconstruct the contour of the defect,and to obtain the final images.Experiment results showed that the average accuracy of this method was 94.73%.When the number of stress wave sensors was reduced from 12 to 6,the defect images reconstructed by this method had higher precision and clearer defect edges than traditional imaging methods,with an average precision improvement of 38.76%.关键词
YOLOv5模型/轮廓约束/木材空洞缺陷/应力波层析成像/传感器数量Key words
YOLOv5 model/constrained contour/wood void defects/stress wave tomography/number of stress wave sensors分类
农业科技引用本文复制引用
叶程浩,杜晓晨,鲍宇..一种基于YOLOv5和轮廓约束的木材空洞缺陷应力波层析成像算法[J].木材科学与技术,2023,37(5):61-68,8.基金项目
浙江省公益技术研究计划项目"面向树木内部缺陷的精细化应力波层析成像技术研究"(LGG19F020019). （LGG19F020019）
