摘要
随着信息技术和计算机视觉技术的发展,仓储管理自动化和智能化成为趋势,对仓储物体进行准确检测变得尤为重要。针对仓储环境下的物体检测应用场景,提出一种基于SSD的仓储物体检测算法,实现对仓储环境下的物体智能检测。首先采用VGG16网络进行图像特征提取,然后在仓储物体数据集上进行模型训练,最后通过优化模型参数将训练好的模型应用于仓储物体检测。在创建的仓储物体数据集上训练SSD300和SSD500两种模型,获得的仓储物体检测准确率(mAP)分别为91.83%和94.32%,表明该算法基本实现了仓储物体的准确检测。
With the rapid development of information technology and computer vision technology,automated and intelligent ware. house management becomes a trend. Therefore,it is particularly important to detect warehouse objects accurately. Aiming at object de. tection application scenario based on warehouse environment,this paper proposes an SSD-based warehouse object detection algorithm which realizes the intelligent detection of warehouse objects. Firstly,the VGG16 network is used to extract the image features,and then the model is trained on the self-built warehouse object dataset. Finally,the trained-well model is applied in the fields of ware. house object detection by optimizing the model parameters. We trained SSD300 and SSD500 models on the self-built warehouse object dataset,and achieved 91.83% mAP and 94.32% mAP,respectively. The experimental results show that proposed algorithm basically realizes the accurate detection of warehouse objects.
作者
陈亮杰
王飞
王梨
王林
CHEN Liang-jie;WANG Fei;WANG Li;WANG Lin(College of Data Science & Information Engineering,Guizhou Minzu University;College of Humanities & Sciences of Guizhou Minzu University,Guiyang 550025,China)
出处
《软件导刊》
2019年第4期28-31,共4页
Software Guide
基金
贵州省教育厅创新群体重大研究项目(黔教合KY字[2018]018)
贵州民族大学科研基金资助项目(2017YB065)
贵州民族大学人文科技学院基金科研项目(18rwjs016)