针对无源毫米波图像普遍存在分辨率不高的问题,提出一种基于正则化技术的超分辨率重建方法.在具体的成像模型基础上,对获取的无源毫米波图像序列进行基于正则的凸集投影(POCS)处理.将正则化条件作为POCS的一个凸约束,使得到的高分辨率图像在平滑噪声的同时能很好地保护图像的纹理细节.实验结果表明,该方法还可以提高图像的分辨率,是一种较好的无源毫米波超分辨率重建方法.
In order to improve the resolution of image acquired by passive millimeter-wave (PMMW) radiometer, a super-resolution reconstruction method based on regularization was presented. The method based on the idiographic imaging model, delt the PMMW image sequence with regularized projection onto convex sets (POCS). Adding the regularized expression to the convex sets can be effective on keeping the textures and reducing noise simultaneously. The experimental tests show that this method can improve the image resolution, and it is a good method for PMMW images' super-resolution reconstruction.