针对高速铁路接触网棒式绝缘子故障的特点和检测需求,提出一种快速模糊匹配及故障检测方法。通过分析获取的现场接触网悬挂装置图像,采用金字塔近邻平均算法对图像与模板分层;对分层后的图像和模板进行归一化互相关计算,提取峰值点,达到快速模糊匹配的目的;然后提取匹配的目标区域,对其横向灰度奇异值检测;最终对棒式绝缘子不良状态做出判断。其识别结果显示该方法有较好的鲁棒性和识别率。
In view of characteristics of faults and measuring requirements of rod insulators of high-speed rail-ways, the fast fuzzy matching and fault detection method was put forward. Through analysis on catenary suspension images obtained on site and neibourhood average pyramid algorithm was adopted to have the image and template stratified. Normalized cross correlation of the stratified image and template was calculated and peak points were extracted to achieve fast fuzzy matching. Then, the matching area was extracted and its transverse gray singular value was detected. Lastly, the bad state of rod insulator was judged. The method has good ro- bustness and recognition rate.