您的位置:首页  > 论文页面

基于卷积神经网络的药片表面缺陷检测

发表时间:2019-12-31  浏览量:464  下载量:29
全部作者: 杨旭东,周学成,张德军,曹聪
作者单位: 华南农业大学工程学院
摘 要: 药片缺陷检测是保证药品质量的重要手段。基于模式识别的药片表面缺陷检测方法存在检测过程复杂、效率和准确率低等问题。为解决上述问题,使用基于卷积神经网络的药片表面缺陷检测方法。首先采用电荷耦合器件(charge coupled device,CCD)工业相机和光电传感器搭建图像采集系统,以实现药片图像的在线自动采集。然后对药片图像进行预处理,去除药片图像的噪声,提高药片图像质量。最后使用卷积神经网络检测药片是否合格,并进一步检测不合格药片的缺陷类型。实验结果表明,在药片表面缺陷检测的准确率上,基于卷积神经网络的检测方法相较于传统机器视觉方法提高了近10个百分点。
关 键 词: 计算机应用;缺陷检测;深度学习;卷积神经网络;机器视觉
Title: Tablet surface defect detection based on convolutional neural network
Author: YANG Xudong, ZHOU Xuecheng, ZHANG Dejun, CAO Cong
Organization: College of Engineering, South China Agricultural University
Abstract: Tablet defect detection is an important means to ensure the quality of tablets. There are many problems in the method of tablet surface defect detection based on pattern recognition, such as complex detection process, low efficiency and accuracy, etc. In order to solve the above problems, a method based on convolutional neural network is used to detect the surface defects of tablets. Firstly, an image acquisition system is built by using charge coupled device (CCD) industrial camera and photoelectric sensor to realize the online automatic acquisition of tablet images. Then the image of the tablet is pre-processed to remove the noise of the tablet image and to improve the quality of the tablet image. Finally, a convolutional neural network is used to detect the eligibility of the tablets and to further detect the type of defects in the unqualified tablets. The experimental results show that the detection method based on convolutional neural network can be improved by nearly 10 percent on the accuracy of tablet surface defect detection, compared with the traditional machine vision method.
Key words: computer applications; defect detection; deep learning; convolutional neural network; machine vision
发表期数: 2019年12月第6期
引用格式: 杨旭东,周学成,张德军,等. 基于卷积神经网络的药片表面缺陷检测[J]. 中国科技论文在线精品论文,2019,12(6):915-922.
 
0 评论数 0
暂无评论
友情链接