摘要
针对电缆沟巡检对机器人低重心运动形态与沟道探测的工作方式的要求,设计了一种基于机器视觉的仿蛇型结构的机器人,它具有沟道运动与探测分析能力。机器人控制系统以Kendryte K210为核心,完成了机器蛇姿态控制发生与控制系统的硬件设计。基于YOLOv2网络模型,完成本地自分类特征提取模型在主控模块上的部署。最后,实现了在PC机上对巡检机械蛇的控制与信息获取。测试结果表明,所设计的电缆沟视觉巡检机器蛇可有效辅助进行日常电缆沟道巡检的工作,提高了工作效率。
In order to meet the requirements of low gravity center movement and channel detection working mode of the robot for cable trench patrol inspection,a snake-like robot based on machine vision is designed,which has the ability of channel movement and detection analysis in this paper.Kendryte K210 is the core of the robot control system,and the hardware design of robot snake attitude control generation and control system is completed.Based on the YOLOv2 network model,the local self-training feature extraction model is deployed on the master control module.Finally,the control and information acquisition of the inspection mechanical snake on PC are realized.
出处
《工业控制计算机》
2021年第5期113-115,共3页
Industrial Control Computer
基金
中国民航大学创新创业项目资助(202010059022)。
