计算机工程与应用 ›› 2010, Vol. 46 ›› Issue (15): 159-161.DOI: 10.3778/j.issn.1002-8331.2010.15.047

• 图形、图像、模式识别 • 上一篇    下一篇

应用PSO与GA混合算法重建超分辨图像

聂笃宪1,陈鹤峰2,袁利国1   

  1. 1.华南农业大学 理学院,广州 510642
    2.广东工业大学 数学学院,广州 510090
  • 收稿日期:2008-11-17 修回日期:2009-02-26 出版日期:2010-05-21 发布日期:2010-05-21
  • 通讯作者: 聂笃宪

Applying hybrid algorithms based on PSO and GA to reconstruct super-resolution image

NIE Du-xian1,CHEN He-feng2,YUAN Li-guo1   

  1. 1.Department of Science,South China Agricultural University,Guangzhou 510642,China
    2.Department of Mathematics,Guangdong University of Technology,Guangzhou 510090,China
  • Received:2008-11-17 Revised:2009-02-26 Online:2010-05-21 Published:2010-05-21
  • Contact: NIE Du-xian

摘要: 结合粒子群优化算法和遗传算法中的交叉与选择操作,提出了一种混合算法,对提出的混合算法用两个具有多个局部极值的函数进行了测试,测试结果表明混合算法寻优能力优于粒子群优化算法;利用该混合算法对低分辨率图像序列重建出一幅高分辨率图像。实验结果表明,该方法重建图像的视觉效果和信噪比均优于遗传算法与梯度下降算子相结合的混合算法重建图像的效果。

关键词: 超分辨率, 图像重建, 粒子群优化, 遗传算法, 混合算法

Abstract: This paper presents a hybrid algorithm integrated Particle Swarm Optimization(PSO) with crossover and select of Genetic algorithms.The proposed hybrid algorithm is tested by two functions with many peak values.Simulated results show that the method proposed in this paper has better ability of finding the global optimum than the standard PSO algorithm.At the same time,the presented approach is applied to restore high resolution image from multiple low-resolution degraded images.Experimental results indicate that the proposed method surpasses the hybrid algorithm method based on GA and descent gradient operator not only in eye fidelity but also in Peak Signal to Noise Ratio(PSNR).

Key words: super-resolution, image reconstruction, Particle Swarm Optimization(PSO), Genetic Algorithms(GA), hybrid algorithms

中图分类号: