自适应稀疏分解的分层介质参数反演  

Adaptive Sparse Decomposition for Layered Media Parameter Inversion

作　　者：周辉林[1] 江涌涛[1] 陈青玉[1] 王玉皞[1] ZHOU Huilin JIANG Yongtao CHEN Qingyu WANG Yuhao(School of Information Engineering, Nanchang University , Nanchang 330031, China)

机构地区：[1]南昌大学信息工程学院,江西南昌330031

出　　处：《雷达科学与技术》2017年第3期295-300,共6页Radar Science and Technology

基　　金：国家自然科学基金(No.61561034;61261010;41505015);江西省自然科学基金(No.2015BAB207001);江西省科技支撑计划(No.20151BBE50090)

摘　　要：为了提高地下多层介质参数反演结果的精度,结合时间延迟完备字典、稀疏自适应匹配追踪以及基于能量的层剥离技术提出了一种新的分层介质参数反演算法。该算法能够对探地雷达信号进行稀疏分解,估计信号的反射系数和时间延迟,进而实现分层介质的参数反演。为了进一步展现该算法的优势,在几组不同分层介质参数的条件下,将该算法的参数反演结果和传统反卷积算法及层剥离技术进行比较。比对结果表明,该算法克服了层剥离估计精度较低和反卷积算法执行时间长等缺点,具有可接受的执行时间和较高的估计精度。In order to improve the precision of the parameter inversion result of the underground multilayered medium. Rooted on time delay over-complete dictionary, sparsity adaptive matching pursuit （SAMP） and energy-based layer stripping, a new layered media parameter inversion algorithm is proposed in this paper. By the proposed algorithm, the ground penetrating radar signal sparse decomposition can be executed, and the signal reflection coefficient and the time delay also can be estimated. And then, the layered media parameter inversion is accomplished. To demonstrate its superiority furthermore, under several groups of different layered media parameters, the algorithm is compared with the traditional deconvolution and layer stripping by use of their parameter inversion results. The comparisons show that the proposed algorithm has higher estimation accuracy and acceptable execution time and overcomes such defects as low estimation accuracy in layer stripping and long execution time in deconvolution algorithm.

关 键 词：探地雷达 分层介质 自适应稀疏分解 时延过完备字典 参数反演 

分 类 号：TN959[电子电信—信号与信息处理]