基于并行遗传算法的图像超分辨率复原

图像超分辨率复原技术，提供了一种利用低分辨率像机获取高分辨率图像的可能途径。图像超分辨率复原有频域方法和空域方法两类：其中频域方法主要基于频谱解混叠；空域方法又分迭代反投影方法、凸集投影方法、Bayesian估计方法等。为了提高图像超分辨率复原的效率和提高复原图像的质量，提出了一种基于并行遗传算法的图像(序列)超分辨率复原的新框架方法，由于遗传算法采用实值编码方式，且基于岛模型的并行机制也有利于多帧图像信息的融合，因而使得算法直观和高效；同时提出采用其他超分辨率复原方法的迭代形式来充当遗传算法的变异算子，因为它能有效地利用已有方法的优点。最后，借用图像复原的客观评价指标来评价超分辨率复原算法的效果。实验证明，该方法有效可行。

A Parallel Genetic Algorithm for Image Super-resolution Restoration

The technique of image super-resolution restoration makes it possible that high resolution images could be restored from low resolution images recorded by low resolution sensors. super-resolution restoration algorithms may be divided into two classes, particularly frequency domain and spatial domain. All frequency domain approaches made use of the aliasing effect; spatial domain algorithms there are mainly three approaches, i.e. Iterative Backward Projecting(IBP), Projection Onto Convex Sets(POCS) and Bayesian methods. In this paper, a parallel genetic framework algorithm for image (sequence) super-resolution restoration is presented . The parallelism of the real-valued genetic algorithm based on the island model enables better integration of the information of the multiple frame images. Especially with the iterative method of other super-resolution algorithms being the mutation operator, the convergence of the genetic searching in the solution space is fast. The experiments demonstrate that the proposed algorithm is efficient and applicable.