摘要
目标检测的一个重要的应用场景就是对于室内人员进行检测,包括人员的流动,人员的状态等等.提出了一种基于yolov 3的教室人员上课状态的检测方法.首先进行数据的标记,标记之后使用vgg进行初次预分类,预分类之后进行人工筛选,然后将数据交给yolo进行训练,最后再将该训练好的模型,放入视频进行实时的检测.结果表明可以较好的检测出来学生的上课状态.
An important application scenario of target detection is to detect the status of the personnel,including the flow and status of personnel and so on.This paper proposes a detection method based on yolov3 to detect the state of personnel in class.The data is manually marked first.Then vgg is used for pre-classification.After the pre-classification,the data are trained by yolov 3.Finally,the trained model is sued for real-time detection by the video.The result shows that the class status of students can be well detected.
作者
常思远
CHANG Siyuan(Information Management Center,Xuchang University,Xuchang 461000,China)
出处
《许昌学院学报》
CAS
2020年第5期128-132,共5页
Journal of Xuchang University
基金
许昌学院校级科研项目(2020YB44)。
关键词
目标检测
上课状态
深度学习
实时检测
target detection
class status
deep learning
real-time detection
