| 引用本文: | 周可慧,廖志伟,陈琳韬,等.基于双重背景分离及自适应网格的电压致热型设备红外图像状态分析[J].电力系统保护与控制,2019,47(24):123-130.[点击复制] |
| ZHOU Kehui,LIAO Zhiwei,CHEN Lintao,et al.Research on state analysis of voltage-heating equipment based on dual background separation and adaptive meshing of infrared image[J].Power System Protection and Control,2019,47(24):123-130[点击复制] |
|
| 摘要: |
| 针对发热量小、温度变化范围窄的电压致热型设备故障红外图像难以识辨的难题，提出一种基于双重背景分离及自适应网格划分的电压致热型设备红外图像运行状态分析方法。以电流互感器套管红外图像为例，首先提出红外图像双重背景分离及部件区域提取算法，即基于RGB和HSV颜色空间特点，转换红外图像颜色空间，分离出第一层背景区域。接着将Ostu算法与Hough变换相结合，提出基于Hough变换的Ostu改进算法，实现背景的第二重分离，得到部件区域。之后，导入红外图像温度，运用自适应网格方法划分部件区域，获得每一网格平均温度，提取温度特征向量，比较待分析图像与正常图像获取温差值，以此判断设备运行状态。最后，通过案例分析验证算法的有效性。同时，对避雷器本体、电压互感器电容单元、断路器支柱等电压致热型设备进行分析，说明算法具有良好的扩展性。 |
| 关键词: 双重背景分离 电压致热型设备红外诊断 自适应网格划分 Ostu HSV |
| DOI：10.19783/j.cnki.pspc.190202 |
| 投稿时间：2019-02-26修订日期：2019-07-12 |
| 基金项目:国家青年科学基金项目资助(61702192);中国南方电网公司科技项目资助(GDKJQQ20152088) |
|
| Research on state analysis of voltage-heating equipment based on dual background separation and adaptive meshing of infrared image |
| ZHOU Kehui,LIAO Zhiwei,CHEN Lintao,HUANG Jiedong |
| (School of Electric Power, South China University of Technology, Guangzhou 510000, China) |
| Abstract: |
| Aiming at the problem that the infrared image of the voltage-heating fault of power equipment with small heat generation and narrow temperature variation is difficult to distinguish, a method for analyzing the operating state of voltage-heating equipment based on dual background separation and adaptive meshing infrared image processing is proposed. Firstly, taking the infrared image of current transformer casing as an example, the dual infrared background separation and component region extraction algorithm are proposed. That is, based on the RGB and HSV color space characteristics, the infrared image color space is converted and the first layer background region is separated. Then, the Ostu algorithm is combined with the Hough transform, proposing the improved Ostu algorithm based on Hough transform to realize the second separation of the background and obtain the component area. Then it imports infrared image temperature, uses adaptive meshing component area method to obtain average temperature of each mesh and extracts temperature feature vector. Comparing the component temperature feature vector of the image to be analyzed with the normal image, the temperature difference is obtained. In this way, the operating state of the equipment can be judged. Finally, the effectiveness of the proposed method is verified by case analysis. At the same time, the voltage-heating equipment such as the arrester body, voltage transformer capacitor unit and circuit breaker struts are analyzed, which shows that the method has good scalability. This work is supported by National Youth Science Foundation of China (No. 61702192) and Science and Technology Project of China Southern Power Grid Company (No. GDKJQQ20152088). |
| Key words: dual background separation infrared diagnosis of voltage-heating devices adaptive meshing Ostu HSV |
