摘要
行人检测技术的应用十分普遍,包括人工智能的研究、智能监控的应用、智能交通、无人驾驶汽车中对行人的检测、对人体行为进行分析后做出预判等,应用深度学习的方法对行人进行检测就是人工智能发展的一个十分重要的方向。文章主要研究的内容有3部分,对原SSD算法进行了改进,探讨了神经网络的根本组成与特性,将原SSD算法的基础VGG16改为ResNet50,提高了检测速度和精确度。
Pedestrian detection has very common applications,including research on artificial intelligence,application of intelligent monitoring,intelligent transportation,detection of pedestrians in autonomous vehicles,analysis of human behavior and making predictions,etc.The application of deep learning methods for pedestrian detection is a very important direction in the development of artificial intelligence.The main research content of the article consists of three parts.It improves the original SSD algorithm,explores the fundamental composition and characteristics of neural networks,and changes the basic VGG16 of the original SSD algorithm to ResNet50,improving detection speed and accuracy.
作者
郭健
王伟
马壮壮
Guo Jian;Wang Wei;Ma Zhuangzhuang(Jiangsu Golden Shield Detection Technology Co.,Ltd.,Nanjing 210042,China)
出处
《无线互联科技》
2023年第5期120-124,共5页
Wireless Internet Technology
