实时手势与目标检测无人机人机交互系统  

Real-time Gesture and Object Detection Drone Interaction System

作　　者：向云龙 雷德俊 许建国 XIANG Yunlong;LEI Dejun;XU Jianguo(Jingzhou Institute,School of Information Engineering,Jingzhou,China,434020;Yangtze University College of Arts and Sciences,School of Information and Electromechanical Engineering,Jingzhou,China,434020)

机构地区：[1]荆州学院信息工程学院,湖北荆州434020 [2]长江大学文理学院信息与机电工程学院,湖北荆州434020

出　　处：《福建电脑》2024年第9期75-79,共5页Journal of Fujian Computer

基　　金：荆州学院2023年大学生创新创业训练计划项目(No.202313245002);2023年荆州市科技局、长江大学文理学院联合科研基金项目资助。

摘　　要：为了提高人机交互效率,本文研发了一种基于视觉的无人机。通过人体手势控制和目标追踪,系统采用Opencv和halcon算子库对无人机拍摄的图像进行处理,结合目标检测模型YOLOv5和Pytorch等算法优化目标检测的实时性和精确度。测试的结果表明,该系统能够准确识别手势并有效执行目标检测,可以广泛应用于安防监控和城市交通管理等领域。In order to improve the efficiency of human-computer interaction,this article has developed a vision based drone.Through human gesture control and target tracking,the system uses Opencv and Halcon operator libraries to process images captured by drones,and optimizes the real-time and accuracy of target detection by combining object detection models such as YOLOv5 and Pytorch.The test results indicate that the system can accurately recognize gestures and effectively perform object detection,and can be widely applied in fields such as security monitoring and urban traffic management.

关 键 词：手势识别 目标检测 无人机 YOLOv5模型 Halcon算法包 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]