摘要
工件表面裂纹检测是工业表面质量检测的重要部分,而表面质量的好坏又是影响工件整体质量的重要因素,因此工件表面质量检测不可或缺。基于工件表面图像特征,结合机器视觉硬件,基于美国NI公司研制的LabVIEW软件平台和NI Vision视觉工具包,研究了工件表面图像模板匹配、去噪、二次分割、裂纹识别的算法,设计了工件表面裂纹检测系统。采用基于灰度的模板匹配方法提取裂纹有效区,将潜在的裂纹区域与背景部分区分开;调用MATLAB script节点对裂纹图像小波去噪,使裂纹部分变得清晰易辨;结合最大类间方差法和形态学处理提取裂纹的连通域图像,成功将裂纹连通域从背景及干扰粒子中提取出来;最后通过计算连通域的圆形度和长宽比特征实现裂纹的存在判断。多个样本工件试验结果表明,该类工件表面裂纹检测系统的稳定性、准确性均优于传统的人工检测,很好地完成了裂纹检测任务。
Workpiece surface crack detection of is an important part of industrial surface quality detection,while surface quality is an important factor affecting the overall workpiece quality,so the detection of workpiece surface quality is indispensable. Based on the feature of surface crack image and combining with the machine vision hardware, the research on workpiece surface image template matching, denoising, binarized segmentation and crack identification is carried out on the basis of Lab VIE W software and NI Vision visual toolkits developed by National Instruments of America, and the workpiece surface crack detection system is designed. The effective area of crack is extracted by adopting the template matching method based on grayscale, and the potential cracked region is distinguished from the background. The wavelet denoising of the crack image is done by applying MATLAB script,to make the crack part clearly and easily to be found. Then the connected domain of crack is extracted by combining the methods of maximum interclass variance method and morphological processing method, to successfully distinguish connected domain of crack from background and interfering particle. Finally, the features of circular degree and length - width ratio of connected domain is calculated to judge the existence of the crack. The results of a series of tests indicate that this kind of workpiece surface crack detection system is better than traditional manual detection on stability and precision, for well accomplishing the crack detection tasks.
出处
《自动化仪表》
CAS
2017年第9期85-88,94,共5页
Process Automation Instrumentation