基于改进SLIC算法的电力设备故障区域分割方法

在电力设备状态监测中,红外测温图像故障区域的分割是今后故障诊断智能化发展的关键环节。为了实现图像自动化处理,提高故障区域的分割精度,提出一种改进SLIC算法的故障区域分割方法。采用导向滤波器对红外测温图像进行预处理;在SLIC超像素迭代过程中增加亮度相似性限制条件,并将生成的超像素以种子点的色彩值匹配颜色;通过自动设置色调阈值,实现对电力设备故障区域的分割和标记。实验结果表明,改进的算法与原始SLIC算法相比,边缘召回率提高了4.10%,对故障区域的分割更具优势。

POWER EQUIPMENT FAULT AREA SEGMENTATION METHOD BASED ON IMPROVED SLIC ALGORITHM

In the maintenance and fault monitoring of power equipment,the segmentation of the fault area in the infrared temperature measurement image is the key link for the intelligent development of fault diagnosis in the future.In order to realize the automatic processing of images and improve the segmentation precision of fault regions,an improved simple linear iterative clustering(SLIC)fault region segmentation method is proposed.It used a steering filter to preprocess the infrared temperature measurement image;the brightness similarity restriction condition was increased in the SLIC superpixel iteration process,and the generated super pixel was matched with the color value of the seed point;by automatically setting the hue thresholds,we relized to segment and mark power equipment failure area.The experimental results show that compared with the original SLIC algorithm,the improved algorithm has an edge recall rate of 4.10%,which is more advantageous for segmentation of fault areas.