基于机器视觉的小模数塑料齿轮的在线检测  

Small Module Plastic Gear On-Line Detection Based on Machine Vision

作　　者：徐俊成[1] 张莉彦[1] 阎华[1] 

机构地区：[1]北京化工大学机电工程学院,北京100029

出　　处：《组合机床与自动化加工技术》2011年第3期57-60,65,共5页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：国家科技支撑计划(2007BAF13B01)

摘　　要：针对塑料齿轮在加工过程中存在的制件不满、飞边等表面缺陷,采用基于机器视觉的非接触检测方法,搭建了完整的齿轮加工质量在线检测系统。研究比较了三种不同的图像去噪处理算法及基于基本全局门限的二值化算法;对检测系统,主要是摄像机及机器人进行了完整标定;在完成齿轮中心提取、边缘轮廓提取后,采用边界跟踪法对轴孔直径进行测量、虚拟圆扫描法对齿形进行检测,最终达到齿轮表面缺陷检测的目的。实践证明,该检测方法检测全面、效率高、且具有较高的检测精度。Aiming at the surface defects of the plastic gear such as part shortage,flashing,the detecting methods of non-contact based on machine vision is adopted and a complete on-line detection system is built.Three different image de-noising algorithms and the basic situation of the threshold detection method are studied.The camera and robot are calibrated.After gear centers extraction,edge extraction,by means of boundary tracking method for measure the diameter of plastic gear,and virtual round of tooth profile for measure tooth profile,the detection of gear surface defects is achieved.Practice proves that the detection method is of comprehensive,high efficiency,and high precision.

关 键 词：机器视觉 直径测量 齿形检测 

分 类 号：TH16[机械工程—机械制造及自动化] TG65[金属学及工艺—金属切削加工及机床]