电机故障检测的小波分析红外图像增强

Using wavelet to enhance infrared image in motor fault detection

针对红外成像电力机车电机检测技术中图像特征不明显,故障点获取困难等问题,提出基于小波分析的红外图像增强算法。采用对图像边缘高频图像信号进行提取,舍去其他高频信号,并对低频分量进行直方图均衡化处理,以此来重新构建红外图像,达到对红外图像去噪、边缘以及故障点增强的作用。通过实验证明,该方法能有效去除图像高频噪声,保留红外图像的边缘特征,对红外图像故障检测提供有效信息。

A new method based on wavelet is proposed to enhance the features of infrared imaging in electric locomotive motor detection. When infrared image decomposed by wavelet, according to the edge of this image, pixels of edge in high-frequency images are retained and the other pixels are zero. And it processes the low frequency image using histogram equalization. With the new low and high images, an enhanced infrared image is reconstructed. Experimental results illustrate that the new method can effectively remove the image noise, retain the infrared image edge features, and provide effective information for fault detec- tion of motor with infrared image.