结合以峭度为稀疏标准的稀疏编码算法的高阶统计特性以及轮廓波分解的方向性和能量变化特性,提出了一种新的基于轮廓波和稀疏编码收缩技术的毫米波图像消噪方法。稀疏编码是一种有效的模拟视觉系统的图像特征提取方法,根据提出的特征系数的稀疏先验分布知识,能够自适应地确定收缩阈值。把该收缩技术应用到轮廓波变换域,能够很好地减弱毫米波图像中的未知噪声。采用相对信噪比评判消噪图像的质量,仿真实验表明,与标准稀疏编码收缩方法、轮廓波变换域降噪方法以及小波软阈值收缩方法相比,该降噪方法能够获得较好的图像恢复质量。
Abstract
Combined the high-order statistical property of the sparse coding, which is based on kurtosis measurement (KSC) and the property of the contourlet's composing orientation and the energy variation, a new denoising method of millimeter wave image, which is based on contourlet and KSC shrinkage technique, is proposed. Kurtosis based Sparse Coding algorithm is an efficient image feature extraction method, which can model the human primary visual system. According to the sparse prior distribution knowledge of feature coefficients extracted, the shrinkage threshold can be determinate. Using this shrinkage technique in the contourlet transform field, the unknown noise contained in millimeter wave image can be reduced efficiently. And utilizing the relative single noise ratio criterion to measure the quality of the image denoised, the simulation experimental results show that comparing with other denoising methods such as sparse coding shrinkage, contourlet denoising and wavelet soft threshold shrinkage, this method proposed here can obtain the better quality of image restoration.
关键词
计量学 /
稀疏编码 /
阈值收缩 /
轮廓变换 /
特征提取 /
图像消噪
Key words
Metrology;Sparse coding /
Threshold shrinkage /
Contourlet transform /
Feature extraction /
Image denoising
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
［1］肖泽龙.毫米波对人体隐匿物品辐射成像研究［D］.南京:南京理工大学,2007.
［2］苏品刚,王宗新,徐正宇,等.毫米波焦平面成像系统［J］.苏州市职业大学学报,2008,19(1):70-73.
［3］范庆辉,李兴国,张光锋,等.阈值法在毫米波目标辐射信号去噪中的应用研究［J］.电子与信息学报,2008,30(10):2356-2359.
［4］Hyvarinen A,Hoyer P,Oja E, et al.Sparse code shrinkage for image denoising［C］//Neural Networks Proceedings,IEEE World Congress on Computational Intelligence,Anchorage,AK,USA, 1998,2:859-864.
［5］尚丽.稀疏编码算法及其应用［D］.合肥:中国科学技术大学,2006.
［6］Do M N,Vetterli M.Contourlets:A directional multire-solution image representation ［C］//Proc of IEEE International Conference on Image Processing,Rochester,New York,USA,2002,1:357-360.
［7］王蕊,尹忠科,龙奕,等.基于改进轮廓波变换的图像去噪算法［J］.计算机工程,2009,35(6):228-230.
［8］张瑾,陈向东.一种基于小波变换的红外图像去噪方法［J］.传感器与微系统,2006,25(8):7-9.
［9］Olshausen B A,Field D J.Emergence of simple-cell receptive field properties by learning a sparse code for natural images［J］. Nature, 1996, 381: 607-609.
［10］Chang S G,Bin Yu,Vetterli M,et al.Adaptive wavelet threshold for image denoising and compression ［J］.IEEE Transaction on Image Processing,2000,9(9):1532-1546.
［11］杨镠,郭宝龙,倪伟,等.基于层结构的Contourlet多阈值去噪算法［J］.计算机工程,2006,32(20):180-182.
［12］Shang Li,Cao Fengwen,Chen Jie,et al.Denoising natural images using sparse coding algorithm based on the kurtosis measurement ［J］.Lecture Notes in Computer Science,2008,5264:351-358.
基金
国家自然科学基金项目（60970058）;江苏省自然科学基金项目（BK2009131）;江苏省“青蓝工程”资助项目;苏州市职业大学创新团队资助项目（3100125)
PDF(1227 KB)
