基于总体最小二乘的Shearlet自适应零值绝缘子红外图像去噪

Faulty Insulators Infrared Thermal Image Adaptive Denoising Based on Total Least Squares Estimation and Shearlet Transform

提出一种基于总体最小二乘(Total Least Square)和Shearlet变换相结合的零值绝缘子红外图像自适应去噪算法。首先利用Shearlet变换对原始图像进行分解,得到各尺度各方向的Shearlet系数;然后考虑到不同尺度Shearlet系数之间的相关性,利用总体最小二乘准确估计了各层的Shearlet系数;最后对处理后的系数进行Shearlet反变换重构得到去噪图像。实验结果表明,该方法与小波去噪法和Bayes估计去噪法相比,能够在有效去噪的同时,更好地保留了图像边界和纹理信息。

Based on the combination of Total Least Squares Estimation and Shearlet Transform, an algorithm for faulty insulators infrared thermal image adaptive denoising is proposed. Firstly, in order to get Shearlet coefficients in all scales and directions, the original image is processed by Shearlet Transform. Then, the correlation of coefficient between different scales of Shearlet taken into account, the Shearlet coefficient in each layer is accurately estimated by using Total Least Squares. Finally, inverse Shearlet Transform is performed to the processed coefficients to reconstruct the denoised image. Experimental results show that compared with Wavelet denoising and Bayes estimate denoising, this algorithm is able to better retain the image edge and texture information while effective denoising.