主动视觉压力容器焊缝表面质量参数检测方法研究进展  

Overview of active vision pressure vessel weld surface quality parameter detection method

作　　者：廖普 刘桂雄[1] 杨宁祥 LIAO Pu;LIU Guixiong;YANG Ningxiang(School of Mechanical and Automotive Engineering,South China University of Technology,Guangzhou 510640,China;Zhuhai Branch,Guangdong institute of Special Equipment Inspection and Research,Zhuhai Guangdong 519000,China)

机构地区：[1]华南理工大学机械与汽车工程学院,广州510640 [2]广东省特种设备检测研究院珠海检测院,广东珠海519000

出　　处：《激光杂志》2021年第7期1-8,共8页Laser Journal

基　　金：国家市场监督管理总局科技计划项目(No.2019MK143)。

摘　　要：压力容器焊缝表面质量参数检测是其安全运行重要保障之一,从视觉检测角度系统评述焊缝表面成像方法、焊缝图像焊缝参数特征点提取方法,重点介绍基于简单激光三角测量焊缝参数检测模型、基于Scheimpflug激光三角测量焊缝参数检测模型,以及阐述基于斜率分析、基于加窗分析、基于曲线拟合、基于角点检测和基于深度学习等5种焊缝参数特征点提取方法,总结分析指出基于Scheimpflug激光三角测量焊缝参数检测模型、基于深度学习焊缝图像参数特征点提取方法等是值得关注的重要研究方向。The detection of surface quality parameters of pressure vessel welds is one of the important guarantees for its safe operation.This article systematically reviews weld surface imaging methods and weld image weld parameter feature point extraction methods from the perspective of visual inspection.And focused on the detection model of weld parameters based on simple laser triangulation and Scheimpflug laser triangulation.Also 5 kinds of feature points extraction methods based on slope analysis,windowing analysis,curve fitting,corner detection and deep learning were described.Finally,we pointed out that the detection model of weld parameters based on Scheimpflug laser triangulation and the feature point extraction method of weld image parameters based on deep learning were important directions worthy of attention.

关 键 词：机器视觉 表面成像 特征提取 深度学习 

分 类 号：TN249[电子电信—物理电子学]