摘要

针对在基于迭代张量高阶奇异值分解(HOSVD)实现运动目标提取过程中面临的正则化参数的手动选择问题, 采用Morozov's偏差准则的方法实现基于HOSVD的运动目标检测的正则化参数自适应估计。正则化参数根据误差水平进行选择和调整, 在算法迭代中实现收敛。实验证明, 所提方法减少了调试时间, 并且能较准确完整地提取运动目标。

Abstract

To solve the problem of manual selection of regularized parameters in the process of moving target extraction based on iterative tensor High-Order Singular Value Decomposition (HOSVD), Morozov's deviation criterion is used to achieve the adaptive estimation of normalized parameters in moving target detection based on HOSVD. The regularized parameter is selected and adjusted according to the error level, and rapid convergence is realized in the iterative process of the algorithm. Experiments prove that this method greatly reduces the debugging time, and can accurately and completely extract the moving targets.

补充资料

中图分类号：TP391

DOI：10.3969/j.issn.1671-637x.2018.11.015

所属栏目：工程应用

收稿日期：2017-10-20

修改稿日期：2017-12-07

网络出版日期：--

作者单位    点击查看

备注：杨瑞峰(1994 —),男, 河南信阳人, 硕士生, 研究方向为数字图像处理。

【1】BOUWMANS T, ZAHZAH E H.Robust PCA via principal component pursuit:a review for a comparative evaluation in video surveillance［J］.Computer Vision & Image Understanding, 2014, 122(4):22-34.

【2】LANG S N, ZHAO B, LIU X J, et al.A new image processing method for discriminating internal layers from radio echo sounding data of ice sheets via a combined robust principal component analysis and total variation approach［J］.Science China Technological Sciences, 2014, 57(4):838-846.

【3】JOHN W, ARVIND G, SHANKAR R, et al.Robust principal component analysis:exact recovery of corrupted low-rank matrices by convex optimization［C］//Proceedings of Neural Information Processing Systems, Whistler:MIT Press, 2009:126-143.

【4】XU Y Y.Fast algorithms for higher-order singular value decomposition from incomplete data［J］.Journal of Computational Mathematics, 2017, 35(4):397-422.

【5】ZHANG X J, JAVIDI B, NG M K, et al.Automatic regularization parameter selection by generalized cross-validation for total variational poisson noise removal［J］.Applied Optics, 2017, 56(9):47-56.

【6】XU Y B, PEI Y, DONG F.An extended L-curve method for choosing a regularization parameter in electrical resistance tomography［J］.Measurement Science & Technology, 2016, 27(9):63-74.

【7】姚智颖,刘冬,刘光斌.具有新目标函数的改进型变分贝叶斯［J］.电光与控制,2011,18(8):45-46, 55.

【8】FERMíN S, VILOCHE B.Simple and efficient determination of the Tikhonov regularization parameter chosen by the generalized discrepancy principle for discrete ill-posed problems［J］.Journal of Scientific Computing, 2015, 63(1):163-184.

【9】DONOHO D L, JOHNSTONE J M.Ideal spatial adaptation by wavelet shrinkage［J］.Biometrika, 1994, 81(3):425-455.

【10】张义昕,张海,吴平.改进的二阶互差分测量噪声方差估计算法［J］.电光与控制, 2016,23(8):13-17,38.

【11】WU G, LUO S.Adaptive fixed-point iterative shrinkage/thresholding algorithm for MR imaging reconstruction using compressed sensing［J］.Magnetic Resonance Imaging, 2014, 32(4):372-378.

【12】RAJWADE A, RANGARAJAN A, BANERJEE A.Image denoising using the higher order singular value decomposition［J］.IEEE Transactions on Pattern Analysis & Machine Intelligence, 2013, 35(4):849-862.

【13】刘刚,周珩,王忠勇,等.空域-小波域联合滤波的红外复杂背景抑制［J］.电光与控制,2015,22(5):37-40,45.

【14】徐联微,杨晓梅.基于迭代张量高阶奇异值分解的运动目标检测［J］.计算机应用研究, 2016,33(9):2856-2861.

引用该论文

YANG Ruifeng,HUANG Shan. Moving Target Extraction Based on Adaptive Estimation of Regularized Parameter[J]. Electronics Optics & Control, 2018, 25(11): 79

杨瑞锋,黄山. 基于正则化参数自适应估计的运动目标提取[J]. 电光与控制, 2018, 25(11): 79