基于计算机视觉的药片实时检测系统  

Study on Real-Time Tablet Detection System Based on Computer Vision

作　　者：赵乾[1] 赵硕伟[1] ZHAO Qian;ZHAO Shuowei(Xinjiang Industry Technical College,Urumqi 830021,China)

机构地区：[1]新疆轻工职业技术学院,新疆乌鲁木齐830021

出　　处：《机械与电子》2018年第3期49-51,共3页Machinery & Electronics

摘　　要：为了提升药品缺陷检测效率,降低制药企业的生产成本,设计了一款基于计算机视觉的药片实时检测系统。介绍了计算机视觉检测系统的工作原理和硬件结构组成,并在此硬件结构基础上,提出了一种新型计算机图像分割和特征提取方法。实验结果表明,所述算法和系统能够显著提高药片缺陷检测准确度,可以达到99%,计算机视觉检测系统能够完全满足实时性、高识别率等要求。In order to improve the drug defect detection efficiency and cut down production cost of the pharmaceutical enterprise,this paper designed a real-time tablet detection system based on computer vision.The working principle and hardware structure of the proposed system was introduced,based on which,a novel method for segmenting images and extracting image features was proposed.The experimental results show that the proposed algorithm and system can significantly improve the accuracy of tablet defect detection,reaching 99%,and it can fully meet the requirements of real-time and high recognition rate.

关 键 词：药品缺陷检测 计算机视觉 边缘检测 特征提取 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]