基于双目鱼眼摄像头全景图片的寻人系统  

Human detection based on the panoramic image

作　　者：刘海龙 张智栋 杨停停 LIU Hailong;ZHANG Zhidong;YANG Tingting(College of Information Engineering,Zhengzhou University,Zhengzhou 450001,China)

机构地区：[1]郑州大学信息工程学院,郑州450001

出　　处：《智能计算机与应用》2021年第8期15-18,22,共5页Intelligent Computer and Applications

摘　　要：现如今,常用的服务型机器人在视觉方面仍采用传统的单向摄像进行处理,使其在面对复杂的周围环境时无法做到及时应对。因此,本文提出了一种由全景图片作为机器人视觉信息,并由此驱动底盘进行全向寻人模型。系统采用的外部硬件设备为双目鱼眼摄像头,并较为详细的论述了如何由球型压缩图片转换为最终的全景图片;提取全景图片中的人体位置信息后,向运动底盘发布相应的驱动信息,完成全向寻人的功能。本系统的创新之处在于将全景图片应用于传统的服务型机器人,对未来服务型机器人广阔的应用场景及发展方向提供了一定的创新与实际应用价值。Nowadays,the commonly used service robots still use the traditional one-way camera to deal with the vision,which makes it unable to deal with the complex surrounding environment in a timely manner.Therefore,this research proposes a panoramic image as the robot visual information,which drives the chassis to Omni-directional human search model.The external hardware device of this system is binocular fisheye camera,and discusses in detail how to convert the spherical compressed image into the final panoramic image,extract the human position information in the panoramic image,and then release the corresponding driving information to the moving chassis.The innovation of this system is to apply panoramic images to the traditional service-oriented robot,which provides a certain innovation value and practical application value for the broad application scenarios and development direction of the future service-oriented robot.

关 键 词：全景视觉 图片压缩 人脸检测 底盘驱动 全向寻人 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]