摘要
针对摄像机拍摄运动车辆模糊图像产生振铃效应且图像数据传输量大的问题,提出一种基于压缩感知理论的车辆运动模糊图像恢复方法。该方法先将车辆运动二维模糊图像离散退化小波变换,对高频系数矩阵进行维纳滤波过滤后,采用托普利兹矩阵测量得到高频稀疏系数,有效抑制了振铃,重组高频和低频系数矩阵得到图像高稀疏信号。重构时,采用维纳滤波去噪后进行非线性重构。结果表明,此种方法能大大减少车辆图像传输数据量,以较小误差实现车辆模糊图像的去模糊重构,获取车辆及车牌图像的丰富细节信息。
A restoration method of vehicle motor-blurred images based on Wiener filter compressive sensing is proposed for problems of image ringing effect and data transmission.First,vehicle motorblurred two-dimensional image is transformed for wavelet after discrete,and high frequency sparse coefficient is obtained by the Toeplitz Matrix measurement after Wiener filter,which can effectively suppress the ringing.Second,image sparse signal is obtained by recombination of high frequency coefficients and low frequency coefficients.In the reconstruction,the nonlinear reconstruction is performed after the Wiener filtering.The experimental results show that this method can greatly reduce the amount of vehicle image transmission data,reconstruct vehicle fuzzy images with smaller error,and obtain rich detail information of the vehicle and vehicle license plate images.
出处
《河北北方学院学报（自然科学版）》
2015年第6期6-10,共5页
Journal of Hebei North University:Natural Science Edition
基金
湖南省科技计划项目(2013GK3088)
湖南省哲学社会科学基金项目(11YBA123)
湖南省教育厅教改项目(湘教通[2014]247号-620)
公安部科技创新项目(2013YYCXHNST035)
关键词
压缩感知
运动模糊
维纳滤波
图像恢复
compressive sensing
motion blur
Wienerfilter
image restoration
