一种基于DSP的PET瓶盖缺陷检测算法的设计

A PET Cap Detecting Algorithm Based on ARM and DSP

提出了一种基于DSP的PET瓶盖缺陷检测算法,能快速检测瓶盖内外圆边缘是否有缺陷,密封圈内是否有污点。检测算法主要包括图像预处理、二值化、感兴趣区域提取、瓶盖的边缘提取、缺陷检测等。实验证明:本算法每小时可检测瓶盖数36000个以上,可检测出0.1mm的缺陷。

The paper describes a rapid algorithm to detect the deficiency of plastic bottle cap and the cap sealing ring if it is stained without touch. The algorithm includes image preprocessing, extraction of the interesting region and cap edge, defect detection and so on. The rationality and stability of the online inspection system are proved by theory and test. The algorithm can detect more than 36,000 caps per hour and 0.1mm defects.