Abstract: A defect detection system based on template matching algorithm is proposed and developed to address the issue of missing key plugging in the field of automotive painting production. The system adopts a shape based template matching algorithm and an image pyramid search algorithm, which effectively reduces the impact of illumination and noise while ensuring real-time on-site operation. After the system is integrated into the production line, the operating results show that the system has a detection accuracy of 99.6%, strong resistance to illumination changes, low cost, and good application value.

Key words: Template matching; Missing plugging; Machine vision; Defect detection

摘要： 针对汽车涂装生产领域中关键堵件漏装的问题，文章提出和开发了一套基于模板匹配 算法的堵件漏装的缺陷检测系统。该系统采用基于形状的模板匹配算法和图像金字塔搜索算 法，既有效降低了光照和噪声的影响，又保证了现场运行的实时性。系统并入生产线后，运 行结果显示，系统检测准确性达到 99.6%，并且抗光照变化能力强、成本低，具备良好的应 用价值。

关键词: 模板匹配；堵件漏装；机器视觉；缺陷检测