摘要
基于传统的迭代终止条件,应用匹配追踪(MP)稀疏分解方法分解高噪信号,研究了固定迭代次数及阈值迭代终止条件的不足.引入相邻残差之比,规格化相邻残差之差,减少了噪声的干扰,设计了残差比阈值迭代终止条件.采用残差比阈值迭代终止条件,对高噪声超声信号进行稀疏分解,分解过程会根据噪声成分的水平自适应地终止迭代,克服了传统迭代终止条件无法选择迭代终止阈值的问题.理论推导与实验都证明了残差比阈值迭代终止条件具有很强的鲁棒性.
The residual ratio iteration termination condition for matching pursuit (MP) method was developed. It can erase the disadvantage of the traditional termination condition which can not terminated in terms of an ultrasonic signal with extremely low SNR. The error criterion is used to monitor the acceptability of the unit-norm vector obtained by the matching pursuit performance. Moreover, the matching pursuit operation will continue until the error ratio is less than a predetermine threshold. The error criterion is an important factor for the signal approximation while noise level is unknown. The numerical simulations and experimental results prove the good performance of the proposed approach which is robustness to be as an iteration termination condition for MP method.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2010年第2期171-175,共5页
Journal of Shanghai Jiaotong University
基金
国家自然科学基金资助项目(60702045)
国家高技术研究发展计划(863)资助项目(2005AA602021)
校青年基金项目
关键词
匹配追踪
稀疏分解
迭代
残差比阈值
matching pursuit (MP)
sparse decomposition
iteration
residual ratio