针对现有非下采样轮廓波变换(NSCT)超分辨率重建方法的不足提出了一种改进的重建方法。空间频率大小反映图像细节信息丰富的程度，改进方法将区域窗口内空间频率的大小作为定权的标准对NSCT 分解获得的各对应高频图像进行自适应加权融合。将NSCT 分析与自适应加权融合相结合用以实现影像超分辨率重建，其中利用自适应加权融合方法将各高频图像分别进行融合，同时将低频图像进行取均值处理，分别获得处理后的高低频图像，通过NSCT逆变换获得最终的超分辨图像。通过仿真实验与工程应用验证了改进方法的可行性与有效性。

An improved method is proposed to solve the existing problem that the fusion process is too simple in non-subsampled contourlet transform (NSCT) super-resolution reconstruction. The magnitude of the spatial frequency reflects the degree of image detail richness, and the improved method uses this magnitude of regional window as the standard of the weight, so the adaptive weighted fusion method is applied to the obtained high frequency by NSCT decomposition of image. The adaptive weighted fusion method and NSCT analysis are assembled to achieve image super-resolution reconstruction, in this process, each corresponding high-frequency image is fused by adaptive weighted fusion method, and the low-frequency image is processed to get a mean. The super-resolution image is acquired by inverse NSCT to both the low frequency images and high frequency images. Through simulation and engineering practice, the improved method proves to be feasible and effective.