摘要
针对积木零件种类繁多、人工分拣效率低等问题,提出一种基于改进YOLOv5积木小零件算法检测系统。该算法使用双层Mosaic-16进行数据增强,利用RGB矩阵完成对比度调整,通过对数据集的优化,实现对YOLOv5算法的改进。实验结果表明:改进后的YOLOv5算法能快速准确地对积木小零件进行识别分类,相比原始YOLOv5算法,模型的训练速率和准确率大大提高。
Targeting at the problems of various kinds of building block parts and low manual sorting efficiency,a small part of building block detection algorithm based on improved YOLOv5 is proposed.The algorithm uses double-layer Mosaic-16 for data enhancement and RGB matrix for contrast adjustment.Through the optimization of the dataset,the YOLOv5 algorithm is improved.The experimental results show that the improved YOLOv5 algorithm can quickly and accurately identify and classify small building block parts.Compared with the original YOLOv5 algorithm,the training speed and accuracy of the model are greatly improved.
作者
徐微
郝琦琦
李波波
黄思绒
XU Wei;HAO Qiqi;LI Bobo;HUANG Sirong(Department of Electrical and Information Engineering,City College,Xi’an Jiaotong University,Xi’an Shaanxi 710018,China;Middling Coal Shaanxi Energy and Chemical Group Co.,Ltd.,Yulin Shaanxi 719000,China;Nationl Institute Corporation of Additve Manufacturing,Xi’an Shaanxi 710300,China)
出处
《电子器件》
CAS
2024年第4期1116-1120,共5页
Chinese Journal of Electron Devices
基金
陕西省教育科学“十三五”规划2020年度课题项目(SGH20Y1379)。
