基于改进粒子群算法的剪切散斑相位图去噪方法

空间相移剪切散斑干涉技术具有全场、非接触、高灵敏度等特点，是动态无损检测的关键技术。针对瞬态剪切散斑干涉获得的高噪声相位条纹图中噪声强度大、条纹复杂等情况，常规粒子群优化算法在高噪声相位图像的去噪处理中存在处理不完整、无法较好保持条纹细节等问题，因此提出一种基于优化粒子群算法的剪切散斑相位图去噪方法。该方法在常规粒子群优化算法的基础上，改进了传统线性惯性权值调整方法，提出非线性权值分配方法，同时通过调整聚集度系数提高了算法局部搜索能力。实验结果表明，该方法能够有效地保护条纹的边缘纹理和相位信息，与常规粒子群优化算法相比速度提高了15%,相位奇异点数减少了21.3%,与其他现有方法相比，所提出的算法的去噪效果更好。

Shear Speckle Phase Pattern Denoising Method Based on Improved Particle Swarm Optimization Algorithm

Spatial phase-shifted shear speckle interferometry is a key technology of dynamic nondestructive testing,which has the characteristics of full-field,noncontact and high sensitivity.In view of the large noise intensity and complex fringe in the high noise phase fringe pattern obtained by the transient shear speckle interferometry,the conventional particle swarm optimization (PSO) algorithm has some problems in the denoising processing of the high noise phase image,such as incomplete processing and unable to better maintain the fringe details.Therefore,a shear speckle phase pattern denoising method based on the improved PSO algorithm is proposed.Based on the conventional PSO algorithm,the traditional linear inertia weight adjustment method is improved,a nonlinear weight allocation method is proposed,and the local search ability of the algorithm is improved by adjusting the aggregation coefficient.The experimental results show that the proposed method can effectively protect the edge texture and phase information of the fringe.Compared with the conventional PSO algorithm,the speed is increased by 15% and the number of phase singular points is reduced by 21.3%.Compared with other existing methods,the proposed algorithm has better denoising effect.