文献类型：期刊文献

中文题名：一种刮研机器人及其高点识别算法的设计

英文题名：Design of New Auto-Scraping Robot and Its High-Points Recognition Algorithm

作者：曹恒督[1];高志[1]

机构：[1]华东理工大学机械与动力工程学院,上海200237

年份：2021

期号：10

起止页码：255

中文期刊名：机械设计与制造

外文期刊名：Machinery Design & Manufacture

收录：CSTPCD;;北大核心:【北大核心2020】；

语种：中文

中文关键词：刮研机器人;机械结构;控制系统;标定;高点;分割识别

外文关键词：Auto-Scraping Robot;Mechanical Structure;Control System;Calibration;High-Points;Segmentation and Recognition

摘要：传统的刮研工艺主要是工人手持刮刀依靠人力刮削高点,效率低下且危害身体健康。随着计算机视觉技术与机器人技术的发展,用视觉图像信息控制机器人作业成为一种趋势。设计了一种新型的基于图像信息控制的自动刮研机器人,对其机械结构和控制系统进行了阐述,并对设备所用相机进行了标定与图像鸟瞰化。针对刮研高点之间相互粘连的现象,研究开发了针对平面高点的分割识别图像处理算法,实际图像处理实验显示该算法可以准确的定位刮研高点,达到使用高点的位置信息来控制直角坐标机器人刮削作业的目的。
The traditional scraping technology is mainly done manually by workers holding a scraper to skive the high-points,which is low efficient and harmful to health in practice.As the computer vision and robot technologies have been developing dramatically,it has become a trend to use image information in robot controlling.A new kind of auto-scraping robot is designed in this paper based on image information control method,and its mechanical structure and control system are introduced,meanwhile,the calibration of its camera and the bird’s-eye view of its image are done.In order to make a distinction between different high-points which are connected to each other,a new segmentation and recognition image processing algorithm is developed.According to the results of image processes of scraping planes,the high accuracy of locating the different high-points is verified.