摘要
针对稀疏分解过完备字典中原子数量庞大导致稀疏分解计算复杂的缺点，文中提出了一种基于改进的Gabor原子结构的快速稀疏分解。该稀疏分解根据Gabor原子本身的特点，通过理论推导，证明了位移因子和频率因子的变化会导致Gabor原子本身相位的变化，其变化范围包含了Gabor原子相位因子的变化。所以它的相位因子在稀疏分解中可以不用考虑，从而减少了计算量，提高了稀疏分解速度。仿真结果表明，对比采用没有去掉相位因子的Gabor原子构建的过完备字典，基于改进的Gabor原子结构的稀疏分解速度提高了11．7倍，并且该算法没有智能计算的随机性缺陷。
Based on the improving structure of Gabor, a fast sparse decomposition algorithm is presented in this paper. According to the characteristic of Gabor, the phase of Gabor atom is varied with the variations of shift factor and frequency factor by theoretical derivation, and the range of the variations includes the variation of phase factor in Gabor atom. Therefore the phase factor is not need to be used during sparse decomposition, then the computation complexity is decreased, and sparse decomposition is quickened. Simu- lation results show that the computational efficiency of the proposed algorithm is about 11.7 times as high as that of using Gabor atoms that its phase factor is not gotten rid of, and the algorithm has not the defect of certain randomness from intelligent computation.
出处
《现代雷达》
CSCD
北大核心
2018年第2期49-52,共4页
Modern Radar
基金
重庆市教委科学技术研究项目(KJ1602909,KJ1503004)
国家自然科学基金资助项目(61371164)
重庆电子工程职业学院智能机器人技术研究中心资助课题(XJPT201705)
关键词
稀疏分解
Gabor原子
快速算法
匹配追踪
过完备字典
sparse decomposition
Gabor atom
fast algorithm
matching pursuit
over-complete dictionary
