基于倒谱特性的运动模糊图像PSF参数估计  

Cepstral feature based motion blur image PSF parameter estimation

作　　者：王秋云[1,2] 

机构地区：[1]甘肃政法学院公安技术学院,甘肃甘肃兰州730070 [2]甘肃省证据科学技术研究与应用重点实验室,甘肃兰州730070

出　　处：《自动化与仪器仪表》2014年第12期180-183,共4页Automation & Instrumentation

基　　金：甘肃政法学院公安视听技术创新教学团队项目甘肃政法学院重点科研项目(GZF2012XZD2203)

摘　　要：根据运动模糊图像的特点,提出一种基于倒谱的运动模糊图像PSF参数估计方法。首先,分析了运动模糊图像退化模型以及运动模糊图像频谱和倒谱的特征,对倒谱图像信息滤波增强后,用Radon变换法计算出运动模糊方向,由倒谱中两个负峰值的位置计算出模糊长度。最后,用Lucy-Richardson算法对估计值进行图像复原验证。实验结果表明,该方法能够较精确地估算出运动模糊参数PSF值。According to the characteristics of the motion-blurred images, in this paper gives a method of estimating the PSF parameters of motion-blur images more accurately which base on characteristics of Cepstrum. First of all, this article analyzes degradation model of the motion-blurred images and the characteristics of the motion blurred image spectrum and cepstrum. After that, enhancement of the cepstrum image information by filtering, and calculate the motion blur direction by Radon transform method, and calculating the motion blur length by the two negative peak in the cepstrum. Finally, restoration of motion blur image by Lucy-Richardson algorithm of validation of the estimated value. The experimental results show that, this method can accurately estimate the motion blur parameter PSF value.

关 键 词：运动模糊图像 退化模型 倒频谱特性 滤波 RADON变换 Lucy-Richardson算法 

分 类 号：TN911[电子电信—通信与信息系统]