摘要
文章基于双目视觉识别技术设计了一套激光除草装置的控制系统。该系统采用激光烧灼的方式快速准确地去除田间杂草。控制系统主要由双目相机、激光器、单片机、振镜系统4个部分组成。该研究主要通过改进的YOLOv8视觉识别算法检测杂草坐标位置,控制振镜系统调整激光的射出角度。该系统通过控制激光灼烧杂草嫩芽,抑制杂草生长,实现精准除草的目的。实验结果表明,改进后的杂草识别模型准确率达97.3%,系统响应时间≤150 ms,单株除草时间约250 ms。文章所设计的控制系统运行稳定、杂草识别准确,具有高效的除草效果。
This paper designs a control system for laser weed elimination devices based on binocular vision recognition technology.The system employs laser ablation to rapidly and accurately remove weeds in fields.The control system primarily consists of four components:a binocular camera,laser emitter,microcontroller,and galvanometer system.The study focuses on detecting weed coordinates through an improved YOLOv8 visual recognition algorithm,controlling the galvanometer system to adjust the laser emission angle,and inhibiting weed growth by ablating weed sprouts with the laser,thereby achieving precise weed elimination.Experimental results indicate that the improved weed recognition model achieves an accuracy rate of 97.3%,with a system response time≤150 ms and a single-weed elimination time of approximately 250 ms.The designed system demonstrates stable operation,accurate weed identification,and highly efficient weed elimination performance.
作者
彭炫
戎劲羽
杜震宇
PENG Xuan;RONG Jinyu;DU Zhenyu(Xinjiang University,Urumqi 830047,China)
出处
《无线互联科技》
2025年第10期26-29,共4页
Wireless Internet Science and Technology
基金
2024年国家级大学生创新训练计划项目,项目名称:基于目标检测的智能田间激光除草机器人,项目编号:202410755039。
关键词
视觉识别
激光除草
单片机
双目相机
振镜
visual recognition
laser weeding
microcontroller
binocular camera
galvanometer
