互感器检定系统中外观检测问题的研究

发布时间：2018-10-11 17:31

【摘要】：近年来,国家电网公司在不遗余力地推行计量体系自动化系统的建设。目前已有十多个省市实施了低压电流互感器自动化检定系统,建立集仓库存储、校验计量设备、物流配送为一体的标准化检定中心,并将互感器的出入库、传送、接线、检定等工作全都自动化。然而,在系统实际运作中集装箱中的互感器可能会出现一些状况,比如空位、颠倒、倾倒、规格不一致等。对于这些非常规状况,该系统并未做检测,不仅影响了工作效率,还会损害互感器,造成直接经济损失。本文在以上背景下,并在了解了互感器自动化检定系统的情况和机器视觉外观检测系统的基本原理与相关技术资料的基础上,设计了互感器外观检测的总体结构和检测算法,并实现了软件开发。本文首先介绍了电流互感器自动化检定系统的情况和存在的问题,分析出互感器外观检测的功能要求。其次,设计了外观检测的整体结构和工作流程。然后,研究了一些图像处理算法,分别给出了互感器有无检测、颠倒检测、倾倒检测、电流比字符识别的解决方案和算法流程。本文采用色调直方图和SURF图像匹配算法来检测互感器的位置情况,并采用基于字符特征的多步匹配算法识别互感器的电流比字符。在MATLAB环境下完成了互感器外观检测的软件开发,实现了互感器有无检测、颠倒检测、倾倒检测和电流比字符识别。
[Abstract]:In recent years, State Grid Company is sparing no effort to implement the construction of automatic measurement system. At present, more than 10 provinces and cities have implemented an automatic verification system for low-voltage current transformers. They have established a standardized verification center that integrates warehouse storage, calibration and metering equipment, and logistics distribution. Verification and other work are all automated. However, in the actual operation of the system, the transformer in the container may have some conditions, such as vacant, upside down, toppling, inconsistent specifications, etc. For these unconventional conditions, the system has not been tested, which not only affects the working efficiency, but also damages the transformer and causes direct economic losses. Based on the above background, and on the basis of understanding the situation of transformer automatic verification system and the basic principle and related technical data of the machine vision appearance inspection system, this paper designs the overall structure and detection algorithm of the appearance detection of transformer. And realized the software development. This paper first introduces the current transformer automatic verification system and the existing problems, and analyzes the functional requirements of the appearance detection of the transformer. Secondly, the whole structure and workflow of appearance inspection are designed. Then, some image processing algorithms are studied, and the solution and algorithm flow of transformer detection, reverse detection, tipping detection and current-current ratio character recognition are given respectively. In this paper, hue histogram and SURF image matching algorithm are used to detect the position of transformer, and the multi-step matching algorithm based on character feature is used to identify the current ratio character of transformer. The software development of appearance detection of transformer is completed under the environment of MATLAB, and the detection of transformer, reverse detection, toppling detection and character recognition of current ratio are realized.
【学位授予单位】：华中科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP391.41

【参考文献】