一种改进的图像恢复遗传算法

Improved Genetic Algorithm for Image Restoration

遗传算法作为一种高度并行、随机、自适应搜索算法,比其他传统的搜索算法在全局搜索时更有效．但利用简单遗传算法进行图像恢复时,存在严重“过早收敛”的问题,使得图像恢复质量不理想．设计了一种新的二维染色体编码方法,并与模拟退火算法相结合,实验结果表明,该方法能较好克服“过早收敛”现象,对提高退化图像恢复质量有较好效果．

As a sort of parallel, stochastic and adaptive method for global search, genetic algorithms gain higher computational efficiency than traditional methods. This paper presents a new genetic algorithm for image restoration by incorporating new chromosome coding method and simulated annealing. Experimental results justify our algorithm in terms of the quality of the restored image and the computational complexity.