基于改进yolov5的悬轨机器人控制系统设计与实现
Design and implementation of suspension robot control system based on improved Yolov5
余子衡 1邓宏涛 1李巍1
作者信息
- 1. 江汉大学人工智能学院,湖北武汉 430056
- 折叠
摘要
近年来,随着深度学习技术的不断发展,计算机视觉技术已成为了 一个极其重要的研究热点.文章针对现有计算机视觉系统中成本与算法可用性的平衡问题,设计了 一款以树莓派4b为计算平台的悬轨机器人控制系统.该系统应用改进后的yolov5算法模型,实现对行人目标的识别检测.通过改进后的模型,在以人脸为识别 目标时,精度可达到 89％.
Abstract
As deep learning technology continues to develop,more and more intelligent applica-tions are emerging.Hardware devices used for training and inference are typically based on GPUs,which are expensive and energy-intensive in practical applications.This paper addresses the issue of balancing cost and algorithm availability in existing deep learning systems by designing a rail-mounted robot control system with a Raspberry Pi 4b as the computing platform.The system ap-plies the improved Yolov5 algorithm model to realize the recognition and detection of pedestrian tar-gets.The improved model can achieve an accuracy of 89％when the target is a face.
关键词
深度学习/嵌入式系统/传感器Key words
deep learning/embedded systems/sensor引用本文复制引用
出版年
2024
NETL
NSTL
万方数据