【摘要】 传统图像插值方法和基于字典学习的图像超分辨率重建方法在图像处理速度和重建图像的视觉质量之间很难保证有效的平衡,针对这一问题,文章利用自然图像非局部相似性的特点和GPU(Graphic Processing Unit)并行处理的优势,提出了一种基于CUDA(Compute Unified Device Architecture)的遥感图像快速超分辨率重建算法。实验结果表明,文章提出的方法能够很好的保持图像的边缘及结构信息,获得良好的视觉效果,同时实现遥感图像的快速超分辨率重建。更多还原

【Abstract】 Aiming at problem that it is very difficult to ensure effective balance between image processing speed and visual quality of reconstruction image by using the traditional image interpolation method and super resolution image reconstruction method based on dictionary learning,a fast remote sensing image super resolution reconstruction algorithm based on CUDA( Computer Unified Device Architecture) is proposed by using characteristics of natural images non- local similarity and advantages of GPU’s( Graphic Processing unit) parallel processing. The experimental results show that this algorithm can preserve the image’s edge and structure information to get a good visual effect and achieve fast super resolution reconstruction of remote sensing image.更多还原