基于总变分规整化的湍流退化图像复原RL算法

Restoration of Turbulence-Degraded Images Based on RL Algorithm with Total Variation Regularization

根据RL(Richardson-Lucy)不需要点扩散函数尺寸先验信息的优点和湍流退化模式的特点,将RL图像复原算法应用到气动光学效应退化图像的复原中.针对RL算法极易引起噪声放大和过度平滑的缺点,提出了基于总变分最小化的RL图像盲复原算法.仿真结果表明,该算法在噪声抑制和细节保存方面都优于非规整化的RL算法.

According to the characters of the turbulence-degraded images. Richardson-Lucy（RL） algorithm was introduced to restore aero-optics degraded images. RL algorithm is totally blind, and it does not need the information about the size of the Point-Spread Function （PSF）. However this method tends to amplify the noise and result in excessive smoothing. Here, the RL algorithm with a regularizing constraint based on total variation was proposed. The simulation results show that this method can effectively remove the noise and preserve the edges of the image.