盲环境下稀疏编码监控视频图像降噪仿真  

Sparse Coding Monitoring Video Image Noise Reduction Simulation in Blind Environment

作　　者：陈昭[1] 赵苏艳[1] CHEN Zhao;ZHAO Su-yan(Jilin Police Academy,Experimental Center,Changchun Jilin 130117,China)

机构地区：[1]吉林警察学院实验中心,吉林长春130117

出　　处：《计算机仿真》2020年第3期368-371,共4页Computer Simulation

基　　金：国家自然科学基金(41271435)。

摘　　要：针对盲环境监控视频图像降噪问题,以及当前图像降噪方法中存在的运行效率较低、降噪图像失真度较高等不足之处,结合稀疏编码技术,提出盲环境下稀疏编码监控视频图像降噪方法。根据稀疏表示理论,将其扩展应用到监控视频图像中,利用正交匹配追踪算法对待处理图像进行稀疏编码;采用自适应方式从含噪图像块样本中获取字典,结合自变量分解及拉格朗日算法进行相关问题求解,并据此对图像稀疏编码系数进行优化;结合噪声模型与图像系统的观察模型,对待处理图像进行噪声估计,根据全部噪声估计均值进行图像降噪处理。仿真结果表明,所提盲环境下稀疏编码监控视频图像降噪方法的图像降噪效果优于实验对比方法,且降噪处理时间更短,具有较好的鲁棒性。This paper combined with the sparse coding technology to propose a method to reduce the noise in the sparse coding monitoring video image based on blind environment.According to the theory of sparse representation,it is extended to the monitoring video image and the image to be processed was sparsely encoded by the orthogonal matching pursuit algorithm.Then,the dictionary is obtained from the sample of the noisy image block in the adaptive way.Combined the independent variable decomposition with Lagrange algorithm,relevant problems could be solved.On this basis,the sparse coding coefficient of image was optimized.After that,the noise model was combined with the observation model of image system to estimate the noise in image to be processed.Finally,the noise reduction of image was performed by the mean of all noise estimation.Simulation results show that the denoising effect of the sparse coding monitoring video image in blind environment is better than that of the comparison method.Meanwhile,the noise reduction time is shorter and the robustness is better.

关 键 词：盲环境 稀疏编码 图像降噪 高斯噪声 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]