基于倒谱特性的运动模糊图像复原技术  

Motion blurred image restoration Technology based on cepstrum feature

作　　者：王秋云[1,2] 李雄伟 杨昊亮 WANG Qiuyun;LI Xiongwei;YANG Haoliang(College of Public Security Technology,Gansu Institute of Political Science and Law,Lanzhou,130070,China;Key Laboratory of Evidence Science Research and Application of Gansu Province,Lanzhou,730070,China)

机构地区：[1]甘肃政法学院公安技术学院,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：《电视技术》2018年第12期18-22,共5页Video Engineering

基　　金：甘肃政法学院2012公安视听教学团队;甘肃省证据科学特色学科支持;甘肃省青年科技基金计划(NO.17JR5RA159)

摘　　要：在运动模糊图像的复原过程中,模糊长度和模糊角度的准确估算是决定复原效果的重要环节。本文以运动模糊图像的形成原因和图像退化模型为研究基础,通过傅里叶逆变换获取图像的倒频谱信息,在倒谱域采用Radon算法和负峰值距离测量方法,估算模糊角度和模糊长度,最后通过对复原图像的质量评价来验证模糊参数的估算精度。实验结果表明,该方法可以准确估算运动模糊图像的模糊参数,复原后的图像具有较好的视觉效果,均值和信息熵指标得到提升,验证了基于倒谱特性的复原算法的优越性及可行性。In the restoration process of motion blurred image,the accurate estimation of blurred lengtli and blurred angle is an important link to determine the restoration rffect.Based on the cause of the formationof motion blurred images andthe model of image degradation,thispaperobtains tlie cepstrum information from the inverse Fourier transform.In the cepstrum domain,the Radon al-gorithm and the negative peak distance measurement method are used to estimate the blurred angle and the fuzzy length.Finally,the fuzzy parameters are verified by the quality evaluation of the complex original images.The accuracy of the estimation.The experimental results show that the method can accurately estimate the fuzzy parametersof the motion blurred image.The reconstructed image has good visual effect,and the mean valueand the entropy index are improved.The superiority and feasibility of the restora-tion algorithm based on the cepstrum characteristic are verified.

关 键 词：运动模糊 倒谱特性 Radon算法 模糊长度 模糊角度 

分 类 号：TN751.1[电子电信—电路与系统]