基于机器视觉的汽车线束压接缺陷检测系统

Defect Detection System Design for Automobile Wiring Harness Crimping Using Machine Vision

汽车线束是汽车电路中信号传输的载体,线束的压接是线束整装工艺中至关重要的一环。接触件端子与导线压接的过程中会出现芯线外露、端子变形等质量问题,传统的检测方法是依靠人工肉眼检测,检测效率不高且准确率较低。针对上述问题,设计了基于机器视觉的线束压接缺陷检测系统,利用OpenCV计算机视觉库对图像中线束的特征进行分析,完成线束压接质量的在线检测。实际应用结果表明该视觉检测系统精度达到±0.03 mm,误报率小于3%,漏报率为0,检测时间小于500 ms,满足生产现场实际的应用需求。

The automobile wiring harness is the carrier of signal transmission in automobile circuit,and the crimping of the wiring harness is a crucial part of the wiring harness assembly process.The quality problems such as core exposed and terminal deformation occur frequently on the contact terminal and the wire.The traditional detection method relies on human eye detection which has a low efficiency and precision.To solve the above problems,a defect detection system for automobile wiring harness crimping using machine vision was designed.Open CV computer vision library was used to analyze the feature of the wiring harness in the image,so as to complete the online detection of wiring harness crimping quality.The practical application result shows that the detection accuracy of the vision system is±0.03 mm,whose misdiagnosis rate is less than 3%while missing report rate is zero,and with detection time less than 500 ms,which meets the practical application requirements.