由红外热像仪直接得到的热像会受很多外来因素的干扰而致使图像模糊不清、对比度差, 由此在故障诊断过程中, 需要对热像进行后续增强处理。总结了影响热故障红外检测准确性的外部因素,分析了输电接头热像的基本特征,采用了基于平方根灰度变换直方图修正和全局均衡化两种方法对热像进行了增强处理。在分析上述两种方法优劣的基础上, 提出了一种基于线性平滑滤波的局部均衡化方法。结果表明, 基于线性平滑滤波的局部均衡化方法对于处理输电接头热像更有效。

The thermal images directly obtained from infrared imaging system are usually blurred and contrast is low because of several external factors. So image enhancement is important in fault diagnosis. External factors affecting detection accuracy are summarized. The basic features of thermal images of transmission line joints are analyzed. The thermal images are enhanced using histogram modification based on the square root transformation function and global histogram equalization. In analyzing the merits of the two methods mentioned, a local area histogram equalization algorithm based on linear smoothing filter is presented. Finally, results show that it is a more efficient method for enhancing thermal images of transmission line joints.