CASIA OpenIR  > 毕业生  > 硕士学位论文
Alternative TitleResearch on DSP Based Vision System for Power Transmission Line Inspection Robot
Thesis Advisor梁自泽
Degree Grantor中国科学院研究生院
Place of Conferral中国科学院自动化研究所
Degree Discipline控制理论与控制工程
Keyword巡线机器人 嵌入式系统 Dsp 图像处理 单目视觉 Power Line Inspection Robot Embedded System Dsp Image Processing Monocular Vision System
Abstract为了确保高压输电的安全,定期对输电线及其周围环境进行巡检是非常重要的。为避免繁重的人力巡检,智能巡线机器人成为研究热点。它可以自主沿线行走、跨越障碍并采集线路图像。为了能够自动识别并跨越输电线上安装的各种电力设施,避免碰撞,视觉系统成为巡线机器人最重要的导航系统。 本文在国家863计划的支持下,为高压多分裂输电线路特种巡线机器人开发了一套基于DSP的嵌入式单目视觉导航系统。该系统能够快速的进行图像采集与处理,具有结构紧凑、功耗小等优势,能够在本地运行实时图像处理算法,在线为机器人提供导航信息,不必将图像传回地面基站进行处理,巡线机器人可在不与地面基站联系的情况,实现大范围完全自主越障与行走。 本文为高压输电线路巡线机器人研究并开发出一套基于DSP平台的实时图像处理算法流程。提出了一种高压输电线路障碍物的实时识别算法,算法在DSP平台上得到了实现与验证,能够有效的对线路上安装的间隔棒、防震锤、悬垂线夹、耐张线夹等障碍物进行实时的识别与定位。本文分析了高压多分裂输电线路特有的结构约束特性,提出了基于相线定位的单目测距算法,能准确估算障碍物与巡线机器人的相对距离,解决了在无标定状态下单目摄像机无法准确测距的难题。 基于DSP的软硬件系统及图像处理与障碍物检测、距离估算算法在DSP平台上的运行效果在室内实验室模拟环境及野外真实高压输电线路环境中进行了验证与测试。系统运行良好,算法运行速度快,并具有较高的精度,大大降低了主控制器的负担,能够为机器人主控制器提供有效的线路信息。
Other AbstractIn order to ensure the quality and reliability of electric power supplied to customers, power transmission lines must be inspected regularly. Aim to avoid artificial inspection work, intelligent power line inspection robot which can crawl along the transmission line is developed to accomplish some necessary operations, such as capturing and recording the videos of the environment around the power lines and towers. Several kinds of electric equipments settled on the power transmission lines are acting as obstacles to the inspection robot during continuous inspection work. So vision system is significant for an inspection robot to obtain surrounding environment information and to plan its behavior to stride over obstacles. Supported by 863 National High-Tech Research and Development Plan, we designed an DSP-based embedded monocular vision system to recognize and estimate the distance of the current obstacle for a special kind of power line robot which runs in partite transmission lines environment. This embedded vision system is characterized with compact size, low power cost and excellent performance in image acquisition and processing operations. This vision system can run real-time image processing algorithms on-line, and provide real-time obstacle information to the main controller of the robot, which means the robot can process and save the images and autonomously stride over the obstacles without the contact to the ground center. A new algorithm of the obstacle’s recognition is proposed and can recognize the electric equipments installed on the power lines such as spacers, counterweights, suspension clamp and strain clamp with isolator strings. According to the specific characteristic of the partite lines structure, an algorithm for obstacle distance estimation is proposed which can provide precise distance information of the obstacle to the main controller and help the robot to autonomously crawl along the transmission line in a long distance. The algorithm proposed overcomes the difficulty of distance estimation by monocular system. The system we designed was tested both in indoor simulation transmission lines environment and outdoor actual high-voltage transmission lines environment. Experiments showed that this system can work smoothly with high speed and precision, and can greatly reduce the burden of the main controller of the Inspection Robot, by providing significant routing information.
Other Identifier200728014628053
Document Type学位论文
Recommended Citation
GB/T 7714
雷炜阳. 基于DSP的架空输电线路巡线机器人视觉系统研究[D]. 中国科学院自动化研究所. 中国科学院研究生院,2010.
Files in This Item:
File Name/Size DocType Version Access License
CASIA_20072801462805(3661KB) 暂不开放CC BY-NC-SAApplication Full Text
Related Services
Recommend this item
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[雷炜阳]'s Articles
Baidu academic
Similar articles in Baidu academic
[雷炜阳]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[雷炜阳]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.