粒子群算法参数搜索的共反射面元叠加

共反射面元（CRS）叠加借助于菲涅耳带的理念，可以最大限度地提高信噪比。在CRS叠加中，叠加效果取决于波场运动学参数，因此对参数的搜索尤为重要，而一些局部优化算法效果受限于参数初值的选取。本文采用粒子群（PSO）算法实现对二维CRS参数的搜索，在不影响叠加效果的基础上，提高了参数搜索效率，大大缩短了计算时间。模型和实际数据处理结果验证了方法的有效性。

Common reflection surface stack based on parameter searching by PSO

The common reflection surface (CRS) is based on the theory of the Fresnel zone,the signal-to-noise ratio (SNR) can be maximized while maintaining the resolution.CRS stack effect depends on the wavefield kinematic parameters.Therefore,the search for these parameters is particularly important.If using point-by-point calculation or violent search to find the best parameters,the calculation is too difficult to be achieved.Some local optimization algorithms can significantly reduce the amount of computation,but its effectiveness is limited by initial parameters.This thesis attempts to use particle swarm optimization (PSO) to achieve the search of 2D CRS parameters.So the efficiency of CRS parameter search is improved and the calculation time is greatly shortened.Tests on model and real data verify the validity of the proposed method.