Abstract: As for the Simple Genetic Algorithm (SGA) in the image restoration application, a new method was proposed to deal with the problem of low matching degree and different matching values, which could make it difficult to obtain the required solutions. This method sorting the matching value into two types of light and shade from the searching solution space was composed by the SGA and the Rough Adaptive Algorithm (RAA).Then in order to enhance the robustness of image recovery algorithm, the two types were dealt with respectively by the rough adaptive model on the basis of the sorting value. Compared with the inverse filter, Wiener filter and SGA, the proposed method has better image edge and higher PSNR.

Key words: Rough-Adaptive model, Genetic Algorithm (GA), operator optimum, image restoration, fitness function

摘要： 针对简单遗传算法(SGA)在图像恢复应用中寻求匹配近似解时,存在匹配度低及匹配值差异较大,导致很难得到所需近似解的问题,设计了一种新的图像恢复方法。该方法采用的方案是将简单遗传算法与粗糙自适应算法相结合,按照匹配数值对SGA在其搜索解空间所得匹配近似解进行明暗标记分类,然后按照粗糙自适应模型进行相应地分类处理,以增强图像恢复算法的鲁棒性。通过与逆滤波、维纳滤波和简单遗传算法的对比实验表明,粗糙自适应遗传算法(RAGA)能更好地保留图像边缘及提高峰值信噪比值。

关键词: 粗糙自适应模型, 遗传算法, 算子优化, 图像恢复, 适应度函数