摘要
本论文以RISC-V处理器架构的K210为核心芯片,利用K210强大的处理AI和图像处理能力,实现对大学校园、小区等交叉路口上行人和交通工具的实时检测.当通过摄像头检测到相关行人和交通工具等时,输出采用交通灯进行相关提示.本文首先介绍了K210的硬件设计及其开发环境建立,接着详细介绍了训练模型的生成、图像目标识别的实现过程,最后阐明智能交通灯的实际测试效果满足设计要求.
This paper takes the K210 of the RISC-V processor architecture as the core chip,and uses the K210’s powerful processing AI and image processing capabilities to realize real-time detection of pedestrians and vehicles at intersections of university campuses and residential areas.When relevant pedestrians and vehicles are detected through the camera,the output uses traffic lights to indicate the relevant information.This paper first introduces the hardware design of K210 and its development environment establishment,then introduces the generation of the training model and the realization process of image target recognition in detail,and finally clarifies that the actual test results of intelligent traffic lights meet the design requirements.
作者
李卓寰
刘煜
彭程
Li Zhuo-huan;Liu Yu;Peng Cheng(School of Energy and Mechatronics Engineering,Hunan University of Humanities,Science and Technology,Loudi 417000,Hunan Province,China)
出处
《科学与信息化》
2023年第1期160-162,共3页
Technology and Information
基金
湖南人文科技学院大学生创新创业训练计划资助项目,项目名称:基于深度学习的人体识别智能行人保护系统,项目编号:S202010553020。
关键词
K210
智能交通灯
图像目标识别
训练模型
K210
intelligent traffic light
image target recognition
training model
