变电站智能巡检机器人视觉导航方法研究

发布时间：2018-08-01 17:23

【摘要】：变电站是电网的枢纽，定期巡视对于保证变电站电气设备正常运行至关重要。近年来，随着电力科技进步和体制改革的不断深入，变电站巡检技术正由机器人巡检逐步取代人工巡检。变电站智能巡检机器人携带红外热像仪、可见光CCD、拾音器等检测与传感装置，以自主运行或远程遥控的方式，全天候地在无人或少人值守的变电站对室外高压电力设备进行定期巡检，能够及时发现电力设备的热缺陷、异物悬挂等异常现象，自动发出报警信号或进行预先设置好的故障处理措施，可靠的导航是巡检机器人正确工作的关键技术，因此研究变电站智能巡检机器人导航方法具有重要的实际意义。本文通过分析现有变电站巡检技术，引入机器视觉和人工智能等基本理论，对变电站巡检机器人功能需求进行分析，首先设计了一种变电站智能巡检机器人总体结构，然后针对无人值守变电站巡检机器人导航定位的问题，制定了一种基于引导线的单目视觉导航方案，，方案包括循线行走和定点检测两部分。处理循线行走部分时提出一种基于先验模板的四边检测路径导航算法，通过对循迹摄像机采集到的道路图像的四条边界信息进行HSL模型转换、图像预处理等分析，比对模板判断类型从而预测性地控制机器人的运动方向。处理重要设备定点检测部分时对变电站需要巡检的一次设备做了统计，进而设计了一种用QR码加载主要待检设备信息的编码方案，从而实现QR码的辅助巡检定位，且变电站新增监控区域时只需更换QR码，从而避开了整个巡检策略变更的繁杂。论文最后模拟机器人实际导航情况进行实验，程序运行结果表明本文的视觉导航方法可靠、减少了图像处理花费的时间进而提高了路径识别速度，且路径识别效果较理想，具有灵活的可移植性。
[Abstract]:Substation is the hub of power network, and regular inspection is very important to ensure the normal operation of substation electrical equipment. In recent years, with the progress of power science and technology and the deepening of system reform, substation inspection technology is gradually replacing manual inspection by robot inspection. The substation intelligent inspection robot carries infrared thermal imager, visible light CCD, sound pickup and other detection and sensing devices to operate independently or remotely by remote control. Conducting regular inspection of outdoor high-voltage power equipment in unattended or under-manned substations around the clock, which can timely discover abnormal phenomena such as thermal defects and foreign body suspension of power equipment. It is the key technology to automatically send alarm signals or pre-set fault handling measures, and reliable navigation is the key technology for the correct operation of the inspection robot. Therefore, it is of great practical significance to study the navigation method of intelligent inspection robot in substation. By analyzing the existing substation inspection technology and introducing the basic theories such as machine vision and artificial intelligence, this paper analyzes the functional requirements of substation inspection robot. Firstly, a general structure of substation intelligent inspection robot is designed. Then, aiming at the problem of navigation and positioning of unattended substation inspection robot, a single vision navigation scheme based on guide line is developed, which includes two parts: wire-following walking and fixed-point detection. This paper presents a path navigation algorithm based on a priori template for detecting the four edges of road images, which is based on a priori template. Through the HSL model transformation and image preprocessing analysis, the four edge information of the road images collected by the track tracking camera is analyzed by HSL model transformation, image preprocessing, and so on. The motion direction of the robot is controlled predictably by comparing the type of the template. When dealing with the important equipment fixed point detection part, this paper makes statistics on the primary equipment which needs to be inspected in the substation, and then designs a coding scheme of loading the information of the main equipment to be inspected by using QR code, so as to realize the auxiliary inspection and location of QR code. And the new monitoring area of substation only need to change QR code, thus avoiding the complexity of the change of the whole inspection strategy. At last, the simulation of robot navigation is carried out. The program results show that the visual navigation method in this paper is reliable, which reduces the time spent in image processing and improves the speed of path recognition, and the effect of path recognition is satisfactory. Flexible portability.
【学位授予单位】：华北电力大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM63;TP242

【参考文献】