Abstract: This paper deals with the parameter choice problems in the regularization method.A novel approach of obtaining regularization parameter is proposed by making use of particle swarm optimization.Through comparing the proposed method to genetic algorithm by numerical simulation testing and contrasting traditional regularization filtering method with the proposed method by image restoration experiment，experimental results show that the proposed method outperforms the compared methods in solving ill-posed problem and it is an effective method.

Key words: inverse problem, regularization method, regularization parameter, Particle Swarm Optimization（PSO）

摘要： 对正则化方法中正则参数的选择进行了研究，提出了利用粒子群优化算法获取正则参数的方法，通过数值模拟实验，对比了该方法与遗传算法，通过图像恢复实验，比较了传统正则化滤波方法和所提出的方法，实验结果表明，所提出的方法在处理不适定问题时更具有优越性，是一种实用有效的方法。

关键词: 反问题, 正则化方法, 正则化参数, 粒子群优化算法