基于荧光磁粉裂纹探伤的图象识别装置设计  

Design of crack image recognition device based on fluorescence magnetic particle inspection

作　　者：陈浩龙 李擎 郑耿峰 CHEN Hao⁃long;LI Qin;ZHENG Geng⁃feng(Fujian Special Equipment Inspection and Research Institute,Fuzhou 350001,China)

机构地区：[1]福建省特种设备检验研究院

出　　处：《电子设计工程》2020年第2期121-125,共5页Electronic Design Engineering

基　　金：福建省特种设备检验研究院科技计划项目（FJTJ2015009)

摘　　要：针对目前特种设备磁粉探伤主要依赖人工判断,智能化水平低的问题,提出了一种基于荧光磁粉裂纹探伤的图象识别方法,通过对被测工件通磁并喷洒磁悬液,利用紫外线光源照射,采用工业相机采集裂纹照片,最后经软件分析自动识别裂纹图象。为解决荧光磁粉裂纹采集图象存在干扰噪声大关联性强的问题,采用RGB-HSV图象预处理算法和梯度特征提取法识别图象裂纹,设计开发图象识别软件并实验验证算法的可行性。检测人员利用荧光磁粉裂纹探伤图象识别装置可以快速判断工件缺陷问题,提高无损检测效率。Aim at the problem of the current special equipment magnetic particle inspection mainly relies on the low level of artificial detection,an image recognition device based on fluorescence magnetic particle crack detection was proposed.By magnetizing and spraying magnetic suspension on the workpiece under test,irradiating with ultraviolet light source,collecting crack photos with industrial camera,and automatically identifying crack images through software analysis.In order to solve the problem of high interference noise in the image of fluorescent magnetic powder cracks,the image preprocessing algorithm RGB-HSV and gradient feature extraction method were used to identify cracks.The image recognition software was designed and developed,and the feasibility of the algorithm was verified by experiments.Using the fluorescence magnetic particle crack detection image recognition device,the inspector will judge the defects of the workpiece easily and improve the efficiency of nondestructive testing.

关 键 词：无损探伤 荧光磁粉检测 图象处理 特征提取 

分 类 号：TN23[电子电信—物理电子学]