QQ客服 官方微博 反馈留言

题名	110kv输电线路自动巡检机器人及图像处理研究
作者	熊晓明
学位类别	工学博士
答辩日期	2005-04-30
授予单位	中国科学院研究生院
授予地点	中国科学院自动化研究所
导师	谭民 ; 梁自泽
关键词	巡线机器人 视觉导航 阈值分割 Retinex模型 高光检测 Line-inspecting Robot Visual Navigation Thresholding Retinex Model Detection of Spacularities.
其他题名	RESEARCH ON 110KV-POWER-TRANSMISSION -LINE-INSPCETING ROBOT & IMAGE PROCESSING
学位专业	控制理论与控制工程
中文摘要	为保证高压输电线路的安全稳定运行，对输电线路的监控与维护是非常必要的。目前常采用人工方式进行输电线路的检测，这种工作方式工作量较大、危险性较高。随着机器人技术的发展，有必要开发一种能取代人工巡检方式的输电线路自动巡检机器人，用于减轻输电线路巡检的工作量、提高工作效率、降低维护成本和确保输电线路的安全运行。本文主要针对110 高压输电线路自动巡检机器人（以下简称巡线机器人）开展研究工作。 本文首先介绍巡线机器人的国内外研究状况，对一些典型的巡线机器人样机进行了较为详细地描述，并提出了巡线机器人的主要研究方向。 其次，结合高压输电线路的实际路况和障碍物类型，确定了巡线机器人的机构设计方案及其工作原理。并对该机构进行了运动学分析，验证了该机构的合理性及可行性。 第三，针对巡线机器人机构复杂和传感器信息多样的特性，开发了一套控制系统，以实现巡线机器人的自主爬行和越障功能。该控制系统采用分层递阶的体系结构设计，能灵活、可靠地控制巡线机器人的运动。 第四，针对实验室模拟线路上的几种常见障碍物，确定了巡线机器人的导航任务，开发了一套可用于障碍物识别的视觉导航系统。该导航系统综合利用多种图像处理技术与模糊模式识别方法，可较准确、可靠地识别障碍物类型。 第五，根据巡线机器人视觉导航系统对图像分割的要求，对现有的图像分割方法进行了分析与比较，总结了各自的优缺点及适用范围，并选择将适用性较强的最小交叉熵的阈值分割方法用于视觉导航系统中。 第六，针对巡线机器人视觉导航系统中图像的光照影响，探讨了基于Retinex模型的彩色图像增强方法的增强效果，并将灰色关联度引入到该增强图像的颜色逼真度的分析中，提出了一种客观的评价准则。 第七，为检测巡线机器人视觉导航系统中图像上的金属反光面（高光），对现有的高光检测方法进行了分析，提出了一种基于Retinex模型的检测方法，该方法简单实用，对于背景简单的图像有良好的检测效果。 最后对本文所取得的研究成果进行了总结，并对下一步工作提出了展望。
英文摘要	In order to guarantee the safe and steady operation of high-voltage power transmission line, it is necessary to watch and maintenance it. At present manual inspection is usually used to do it. But this work is laborious and dangerous. Now as the development of robot technology, it is necessary to develop line-inspecting robot, which aims at the relief of workload, the improvement of wok efficient, the decrease in cost and the guarantee of the operation of line. The paper focuses on 110kv-power-transmission-line inspecting robot, briefly called as 110kv line-inspecting robot as follows. Firstly, the researches on line-inspecting robot are introduced, especially some typical robot prototypes. And the primary research fields of line-inspecting robot are proposed.Secondly, by considering the actual status of power transmission line, the machine of the 110kv line-inspecting robot and its work principle are decided. And the kinematics of the machine is analyzed, which validates the rationalization and feasibility of it.Thirdly, aiming at the complex machine of line-inspecting robot, a control system with hierarchical architecture for the 110kv line-inspecting robot is designed and developed, in order to realize the creeping, spanning and inspecting functions. It can control the movements of the 110kv line-inspecting robot flexibly and reliably. Fourthly, aiming at several typical obstacles on mock line in our lab, the navigation task of the 110kv line-inspecting robot is decided and a visual navigation system is designed and developed. By using of some image processing methods and fuzzy pattern recognition, the visual navigation system can recognize the types of the obstacles in the lab well Fifthly, According to the requirement of image segmentation in the visual navigation system, the present image segmentation algorithms are analyzed and compared, and the advantages and disadvantages of them are summarized. Then apply minimum cross entropy thresholding, whose applicability is wider, to segment the images in the visual navigation system. Sixthly, aiming at the illumination effect on the images in the visual navigation system, the color image enhancement based on retinex model is discussed. And an impersonal appraisal rule about the color rendition of the enhanced images is proposed by introducing the notion of gray relational degree. Seventhly, in order to detect the glistening regions (spacularities) in the images of the visual navigation system, the present detection algorithms are analyzed. And a new detection algorithm based on retinex model is proposed, which can detect them simply and effectively especially for images in simple background.Finally the primary work of the paper is summarized, and the topics for further research are also addressed.
语种	中文
其他标识符	200218014603178
内容类型	学位论文
源URL	[http://ir.ia.ac.cn/handle/173211/5853]
专题	毕业生_博士学位论文
推荐引用方式 GB/T 7714	熊晓明. 110kv输电线路自动巡检机器人及图像处理研究[D]. 中国科学院自动化研究所. 中国科学院研究生院. 2005.

个性服务

查看访问统计

相关权益政策

暂无数据

收藏/分享

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。