通過分析均勻分布與Cauchy分布的分布機(jī)制，提出了一種改進(jìn)的模擬退火圖像盲復(fù)原算法，該算法選擇Cauchy分布為隨機(jī)擾動量來產(chǎn)生狀態(tài)擾動函數(shù)。通過計算機(jī)仿真，驗證了該算法對初值的魯棒性和復(fù)原的效果優(yōu)于基于均勻分布隨機(jī)擾動量模擬退火盲解卷積算法，提高了收斂到最優(yōu)解的速度。
關(guān) 鍵 詞 圖像復(fù)原; 盲復(fù)原; 盲解卷積; 模擬退火; Cauchy分布

Abstract After analyzing the distribution mechanisms of Uniform distribution and Cauchy distribution, an algorithm of blind image restoration by simulated annealing based on Cauchy distribution random perturbation was proposed. Simulation results show that the improved method has better quality、robustness and speed of convergence comparing with simulated annealing algorithm based on Uniform distribution random perturbation.
Key words image restoration; blind restoration; blind deconvolution; simulated annealing; Cauchy distribution