采用机器视觉方法检测产品时，针对不同产品必须重新编程开发程序，甚至不同规格的同类产品也面临同样的问题。开发了一种可重构的、开放性的智能缺陷检测系统。基于OpenCV/C++联合编程，优化了缺陷检测算法，研发了缺陷区域面积像素值的自定义功能，以供对精度和速度有不同要求的用户进行个性化定制。以手机适配器激光打标缺陷检测为例，设计了缺陷检测系统，检测结果表明，该系统功能稳定，性能可靠，具有较好的实用性。

At present, when using machine vision methods to detect product defects, different products must be reprogrammed and developed, and even similar products with different specifications are facing the same problem. A reconfigurable and open intelligent defect detection system was developed. Based on the OpenCV/C++ joint programming, the defect detection algorithm is optimized, and the custom function of the pixel value of the defect area is developed to meet the needs of users with different precisions and speeds for personalized customization. Taking the defect detection of mobile phone adapter laser marking as an example, the defect detection system is designed. The test results show that the system has stable function, reliable performance and good practicability.