摘要
设计了一个基于四旋翼飞行器的巡线机器人,巡检电力线路及杆塔状态,发现异常时拍摄并存储,任务结束后数据传送到地面显示装置上显示。视觉部分使用3个摄像头,其作用分别为:基于树莓派搭载OpenCV环境,使用稠密光流算法,采集正下方图像,在避免摩尔纹干扰的前提下得到光流数据,进而得到四旋翼飞行器的速度;基于OpenMV,使用霍夫变换算法检测起点圆用于定点起飞,使用MicroPython里的morph函数通过过滤器内核对条纹图像进行卷积得到飞行路径;基于OpenMV,通过色块信息对黑杆距离进行计算以实现绕杆转180°的要求,使用MicroPython里的函数检测条形码和二维码。经测试,3个摄像头协同工作实现了四旋翼飞行器对当前位置和状态的判断,完成电力线路及杆塔状态检测。
A line inspection robot based on quadrotor aircraft is designed to inspect the status of power lines and towers,take pictures and store them when abnormal conditions are found,and transmit them to the ground display device when the task is over.The visual part uses three cameras,whose usages are as follows.Based on the OpenCV environment carried by raspberry Pi,the dense optical flow algorithm is used to collect the image directly below to obtain the optical flow data under the premise of avoiding Moire interference,and the speed of the four rotor aircraft can be obtained.Based on OpenMV,the detection circle is used to detect the circle by Hoff transform,and the starting circle is used for the fixed point takeoff coordinates.The flight path is obtained by convoluting the fringe image by using the morph function in MicroPython.Based on OpenMV,the black bar is detected,and the distance of the black bar is calculated through the color block information to achieve 180 degrees around and using functions in MicroPython to detect barcode and QR code.Through the test,the three cameras work together to realize the judgment of the current position and state of the four rotor aircraft,and complete the function of detecting the status of power line and tower photographing and storing.
作者
余善恩
李真
邓文渝
YU Shanen;LI Zhen;DENG Wenyu(College of Automation,Hangzhou Dianzi University,Hangzhou 310018,China)
出处
《实验室研究与探索》
CAS
北大核心
2022年第1期74-79,共6页
Research and Exploration In Laboratory
基金
国家自然科学基金项目(62073111)
浙江省高等教育“十三五”教学改革研究项目(JG20190172)
杭州电子科技大学高等教育教学改革研究项目(ZDJG201909,YBJG202169)。
关键词
摩尔纹
光流
圆检测
核滤波
单目测距
Moire pattern
optical flow
circle detection
kernel filtering
monocular ranging
