基于深度学习的食用油灌装质量检测系统

摘要

介绍了一种利用深度学习算法进行食用油灌装质量检测的系统,基于深度学习有监督物体识别网络对食用油生产线进行从原料至销售的全流程包装缺陷检测,具体功能包括瓶口缺陷检测、瓶盖缺陷检测、瓶身喷码缺陷检测、贴标缺陷检测、装箱点数检测.相比于传统机器视觉检测方案,该系统具有无需做图像预处理、检测精度高、参数设置简单、算法泛化能力强、开发周期短的优点,可实现食用油生产包装质量检测的全面自动化.%An edible oil filling quality detection system based on deep learning networks is introduced.The algorithm of this system adopts the supervised object recognition networks.The system can realize the whole-production-line detection of filling liquid packaging defects,including defects detection of bottle mouth,bottle cap,printing characters,sticking label,as well as packing quantity.Compared to the traditional machine vision detection,the system has the advantages of free image preprocessing,high detection precision,few control parameters,wide generality of the algorithrms and short development cycle,which can achieve a comprehensive automation of edible oil filling quality detection.