基于视觉的地面污迹清除机器人研究
发布时间：2018-12-06 09:28
【摘要】：移动机器人在军事,医疗,工业等方面的应用已经逐渐成熟,随着科学技术的发展以及人们生活水平的逐渐提高,以机器人代替人类进行高危险、高精度、机械化劳动已成为一种趋势。目前在人们生活的城市环境中充斥着大量地面污迹垃圾,如广告贴画以及宣传单等,并以其难清除、反复出现等难题严重干扰着人们的正常生活和城市的美观。为此本文设计了一种基于视觉摄像头的地面污迹清除机器人。该机器人能够自动寻找地面目标位置并清除。 本文根据地面清洁机器人的实际需求设计了系统的总体结构,主要以机器人结构分析、视觉系统建模、图像采集处理、定位避障算法作为研究的主体。首先,设计机器人系统的总体构成,叙述了各个功能结构模块及其硬件配置;其次通过地面清洁机器人机械本体上的单目摄像头,对三维环境信息进行捕捉,通过视觉模型将三维信息转换为二维图像信息,并传递给主控制芯片,采用动态阈值法进行图像分割,之后选择标定像素点法对图像特征提取,计算地面目标物的实际位置;最后介绍当机器人行进过程中遇到障碍物时,运用以模糊控制理论为基础的避障算法实现机器人的避障。 通过实验表明,本文设计的地面清洁机器人系统通过图像采集、处理以及对特征像素点的标定等步骤能够有效的识别地面污迹信息,并通过主控模块向执行机构发送指令将其清除,该系统具有较好的清洁效果,同时本文所使用的避障算法能够使机器人安全避障。
[Abstract]:With the development of science and technology and the improvement of people's standard of living, mobile robots have been used in military, medical and industrial fields to carry out high danger and high precision instead of human beings. Mechanized labor has become a trend. At present, people live in the urban environment full of a large number of ground dirt garbage, such as advertising stickers and posters, and because of its difficult to remove, recurring and other problems seriously interferes with people's normal life and the beauty of the city. In this paper, a kind of ground stain removal robot based on visual camera is designed. The robot can automatically locate and clear ground targets. In this paper, the overall structure of the system is designed according to the actual needs of the ground cleaning robot. The main body of the research is robot structure analysis, visual system modeling, image acquisition and processing, location and obstacle avoidance algorithm. Firstly, the overall structure of the robot system is designed, and each functional structure module and its hardware configuration are described. Secondly, the 3D environment information is captured through the monocular camera on the mechanical body of the ground cleaning robot, and the 3D information is converted into two-dimensional image information through the visual model and transferred to the main control chip. The dynamic threshold method is used to segment the image, then the calibration pixel method is selected to extract the image features, and the actual position of the ground object is calculated. Finally, when the robot encounters obstacles in the course of moving, the obstacle avoidance algorithm based on fuzzy control theory is used to realize the obstacle avoidance of the robot. The experimental results show that the system can effectively identify the ground smear information through the steps of image acquisition, processing and calibration of feature pixels. The system has good cleaning effect and the obstacle avoidance algorithm used in this paper can make the robot avoid obstacles safely.
【学位授予单位】：哈尔滨理工大学
【学位级别】：硕士
【学位授予年份】：2011
【分类号】：TP242
[Abstract]:With the development of science and technology and the improvement of people's standard of living, mobile robots have been used in military, medical and industrial fields to carry out high danger and high precision instead of human beings. Mechanized labor has become a trend. At present, people live in the urban environment full of a large number of ground dirt garbage, such as advertising stickers and posters, and because of its difficult to remove, recurring and other problems seriously interferes with people's normal life and the beauty of the city. In this paper, a kind of ground stain removal robot based on visual camera is designed. The robot can automatically locate and clear ground targets. In this paper, the overall structure of the system is designed according to the actual needs of the ground cleaning robot. The main body of the research is robot structure analysis, visual system modeling, image acquisition and processing, location and obstacle avoidance algorithm. Firstly, the overall structure of the robot system is designed, and each functional structure module and its hardware configuration are described. Secondly, the 3D environment information is captured through the monocular camera on the mechanical body of the ground cleaning robot, and the 3D information is converted into two-dimensional image information through the visual model and transferred to the main control chip. The dynamic threshold method is used to segment the image, then the calibration pixel method is selected to extract the image features, and the actual position of the ground object is calculated. Finally, when the robot encounters obstacles in the course of moving, the obstacle avoidance algorithm based on fuzzy control theory is used to realize the obstacle avoidance of the robot. The experimental results show that the system can effectively identify the ground smear information through the steps of image acquisition, processing and calibration of feature pixels. The system has good cleaning effect and the obstacle avoidance algorithm used in this paper can make the robot avoid obstacles safely.
【学位授予单位】：哈尔滨理工大学
【学位级别】：硕士
【学位授予年份】：2011
【分类号】：TP242
【参考文献】
相关期刊论文 前10条
1 宗光华;邓鲁华;王巍;;一种鲁棒的室外移动机器人定位方法[J];北京航空航天大学学报;2007年04期
2 胡劲草;;室内自主式移动机器人定位方法研究[J];传动技术;2006年04期
3 张辉;张道勇;何最红;;灰度等级处理中的OSTU动态阈值法研究[J];传感器世界;2008年07期
4 熊光明;赵涛;龚建伟;高峻\，
本文编号：2365820
本文链接：https://www.wllwen.com/wenyilunwen/guanggaoshejilunwen/2365820.html
