混合1范数与全变分约束下的海底目标探测  

Seabed target detection under the constraint of mixed 1-norm and total variation

作　　者：陈筱月 马晓川[1,2] 李璇 裴兴园[1,2] 扶钰斌 CHEN Xiaoyue;MA Xiaochuan;LI Xuan;PEI Xingyuan;FU Yubin(Key Laboratory of Information Technology for AUVs,Institute of Acoustics,Chinese Academy of Sciences,Beijing 100190;School of Electronic,Electrical and Communication Engineering,University of Chinese Academy of Sciences,Beijing 100049)

机构地区：[1]中国科学院水下航行器信息技术重点实验室(声学研究所),北京100190 [2]中国科学院大学电子电气与通信工程学院,北京100049

出　　处：《声学学报》2023年第4期632-645,共14页Acta Acustica

基　　金：中国科学院战略性先导科技专项(XDA22030102)资助。

摘　　要：提出了一种以1范数与全变分的加权和作为正则项的方位估计算法,该算法通过1范数稀疏约束来获得高分辨能力,采用全变分约束来加强目标边缘特征并提高其内部的平滑性,在提高方位分辨率的同时,改善了对方位扩展目标估计性能。在宽带条件下对算法进行了扩展,将接收信号在频域分频后的子频带变换回时域,再在时域上使用窄带算法进行处理,这样避免了时域分段处理造成的时间分辨力的损失。结合了海底探测场景进行了仿真实验验证,结果显示,所提算法的探测结果不仅获得了方位分辨率提升,而且总平方误差小于目前常见的同类算法。因此,在对海底大型方位扩展目标的探测当中,通过在稀疏优化类方位估计方法中加入全变分约束能够保持目标的结构特征,有效提升探测性能。This paper presents an innovative sparse-constrained arrival estimation algorithm using a weighted sum of the 1-norm and total variation constraints as a regularization term.The proposed algorithm achieves high resolution under the 1-norm sparsity constraint and enhances the target's edge features and internal smoothness under the total variation constraint.In addition,the proposed algorithm can improve both the direction of arrival resolution and the estimation performance of the direction of arrival extended targets.Further,the proposed algorithm is extended to the broadband scenes,where the received signal is transformed back to the time domain after being divided into sub-bands in the frequency domain.The obtained time-domain signals are processed by narrowband algorithms.The main advantage of the proposed method is that it avoids the loss of range resolution caused by conventional piecewise processing.The effectiveness of the proposed algorithm is verified by simulation experiments in a seabed exploration scenario.The experimental results show that the proposed algorithm not only can enhance the azimuth resolution but can also reduce the total square error compared to the commonly-used algorithms.Finally,the results demonstrate that adding the total variation constraints to the sparse-constrained direction of the arrival estimation algorithm can preserve the structural properties of targets and effectively improve the detection performance of large azimuth-expanding targets on the seabed.

关 键 词：海底目标探测 方位估计 稀疏优化 全变分 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]