应变片外观缺陷检测方法

An Appearance Defect Detection Method of Strain Gauge

应变片是用于测量应变的关键元件,在实际中应用广泛,其外观缺陷检测在工业生产中具有重要意义。针对应变片图像缺陷检测存在的难点和人工目视检测应变片外观缺陷存在的成本高、检测结果不稳定、效率低等问题,提出了一种自动检测应变片外观缺陷的方法。其中提出了方向Canny边缘检测算法,在准确检测出应变片图像边缘的同时,有效解决了传统Canny边缘检测算法的耗时问题。针对应变片图像灰度分布不均、光照剧烈变化等情况,提出了动态阈值法来分割丝栅区和焊接区。首先,对应变片图像采用双边滤波进行预处理,在去除噪声干扰的同时保留边缘信息;然后,根据方向Canny算子分割出感兴趣区域;其次,根据动态阈值法分割出丝栅区和焊接区;最后,采用形态学操作的方法,检测应变片存在的缺陷。实验结果表明,本文所提检测方法能够准确有效地检测出应变片外观缺陷,算法稳定性高,满足实时性要求;同时系统人工干预少,检测效率高,能实现自动检测。

Strain gauge is a key component used to measure strain,which is widely used in practice,and its appearance defect de-tection is of great significance in industrial production.Aiming at the difficulties in the detection of strain gage image defects and the problems of high cost,unstable detection results and low efficiency of manual visual detection of strain gage appearance defects,an au-tomatic detection method for strain gage appearance defects was proposed.Among them,a directional canny edge detection algorithm was proposed.While accurately detecting the edge of the strain gauge image,it effectively solved the time-consuming problem of the traditional canny edge detection algorithm.A dynamic threshold method was used to segment the wire grid area and welding area.First-ly,bilateral filtering was used to preprocess the strain gauge image,and the edge information was preserved while removing noise inter-ference.Then,the region of interest was segmented according to the direction canny operator.Secondly,the wire grid area and the welding area were segmented according to the dynamic threshold method.Finally,the method of morphological operation was used to detect the defects existing in the strain gauge.The experimental results show that the detection method proposed can accurately and ef-fectively detect the appearance defects of the strain gauge,the algorithm has high stability,and meets the real-time requirements.At the same time,the system has less manual intervention,high detection efficiency,and can realize automatic detection.