In view of the shortcomings that particle swarm optimization is easy to fall into local optima and difficult to solve complex problems, the combination of Gaussian distribution and Cauchy distribution was used in the position updating formula to improve the particle diversity, and Cauchy perturbation was added to the swarm optimal position to further improve its global searching ability. In the experiment, ten benchmark test functions were used to test two proposed modifications in the study, and compared with the four classical particle swarm algorithms, the results show that the proposed algorithm had high solving accuracy and good solving stability, especially in solving complex functions. The proposed algorithm was utilized in spectral reconstruction based on wideband multi-illuminant imaging. The experimental results confirm that comparing with the classical PSO algorithm, the proposed algorithm is good at searching for global optimum especially for complicated engineering problems.