基于OTSU和区域生长的电气设备多点故障分割

Multi Point Fault Segmentation of Electrical Equipment Based on OTSU and Region Growth

针对电气设备的红外图像背景复杂、信噪比低的特点,提出了一种基于OTSU和区域生长相结合的故障分割方法。首先,采用OTSU算法将故障红外图像分割为种子区域和背景区域,去除大面积背景的干扰;然后采用自动选取种子点的区域生长算法完整的分割出故障区域。试验结果表明该算法针对复杂背景的红外图像能有效地消除干扰,并能完整地提取出故障区域。

A method combining OTSU and region growing is proposed for the infrared imaging of electrical equipment with a complicated background and low signal-to-noise ratio. Firstly, the OTSU method is used to divide the faulty infrared image into a seed area and a background area and remove the interference caused by a large area bright background. Thereafter, the region growing algorithm with automatic seed selection is applied to segment the fault area completely. The experimental results show that the algorithm can effectively eliminate interference for the complex background infrared image, and can extract the fault area completely.