基于机器视觉的电子元器件表面缺陷检测

Surface Defect Detection of Electronic Components Based on Machine Vision

在电子元器件的实际生产过程中,由于各个方面因素的影响,电子元器件表面会产生划痕、脏污等表面缺陷,这些缺陷将会严重地影响元器件的质量。目前主要采用人工目检的方式对缺陷进行检测,但这种方式成本高且误差较大。快速、有效地检测金属表面缺陷已成为提高生产效率的关键。复杂的光照条件会使金属表面产生强反射和倒影,缺陷种类多样、边界模糊,给缺陷检测问题带来巨大的挑战。为了解决上述问题,提出了Canny-区域分裂合并算法。经过实验验证,该方法具有更高的精度和良好的鲁棒性,可应用于实际生产中。

In the actual production process of electronic components,due to the influence of various factors,surface defects such as scratches and smudges may occur on the surface of electronic components,and these defects will seriously affect the quality of components.At present,manual visual inspection is mainly used to detect these defects,but this method has high cost and large error.So rapid and effective detection of metal surface defects has become the key to improve production efficiency.Complex illumination conditions will cause strong reflection on the metal surface,and the defect types are diverse and boundary is fuzzy,which brings great challenges to defect detection.In order to solve the above problems,the Canny-region splitting and merging algorithm is proposed.The experiment proves that the method has higher accuracy and good robustness,and can be applied in production.