基于小波域正则化的毫米波图像重构

Based On Wavelet Domain Regularization Of Millimeter-wave Image Reconstruction

毫米波被动成像具有全天时工作能力。与红外、可见光成像相比,其不足之处是分辨率较低,不能完全反映场景与目标的细节情况。采用小波域正则化方法,首先对毫米波图像进行小波域局部噪声方差估计,然后用自适应正则化方法重构超分辨率毫米波图像。毫米波图像处理的实验证明,该方法消噪效果明显,能锐化图像,保持图像细节。

Millimeter wave passive imaging has the ability to work day. But compared with infrared and visible light imaging, one of most important drawback in millimeter-wave passive imaging is the lower resolution, cannot fully reflect the scene and the target＇s detail. It is most important to improve millimeter-wave image resolution for better detecting and identifying the target. This paper, we use wavelet domain regularization method. Conduct millimeter-wave image wavelet domain local noise variance estimation, the latter with adaptive regularization method for reconstruction of high-resolution millimeter wave images. The experimental results show that this method significantly noise cancellation, sharpen the image, while preserving image detail.