基于空间约束联合字典学习的三维冲击波超压场重建  

Three-dimensional Shock Wave Overpressure Field Reconstruction Based on Spatial Constrained Joint Dictionary Learning

作　　者：魏晓曼 李剑[1] 刘晓佳 郭陈莉 展勇忠 刘代劲 WEI Xiaoman;LI Jian;LIU Xiaojia;GUO Chenli;ZHAN Yongzhong;LIU Daijin(Shanxi Key Laboratory of Signal Capturing and Processing,North Central University,Taiyuan 030051,China;Hunan Yunjian Group Co.LTD,Changsha 410000,China)

机构地区：[1]中北大学信息探测与处理山西省重点实验室,山西太原030051 [2]湖南云箭集团有限公司,湖南长沙410000

出　　处：《探测与控制学报》2024年第2期108-114,共7页Journal of Detection & Control

基　　金：国家自然科学基金项目(6227012725)。

摘　　要：针对有限测点条件下冲击波超压场全时空全区域重建的需求,提出一种基于空间约束联合字典学习的三维冲击波超压场重建方法。建立三维走时层析成像模型,利用字典学习对图像的稀疏表征优势和空间约束对图像边缘纹理保真特性,采用交替最小化思想对目标函数进行优化解算,实现大型病态稀疏矩阵的高精度求解。试验结果表明,该方法与代数重建算法(ART)、联合代数重建算法(SART)和期望极大化算法(EM)相比,重建精度更高,全区域重建误差达到了15.81%,在时空场重建过程中具有较高的应用价值。For the demand of full spatial and temporal reconstruction of shock wave overpressure field under limited measurement point conditions,this paper proposed a 3D shock wave overpressure field reconstruction method based on spatial constraint combined with dictionary learning.Firstly,a 3D walk-time laminar imaging model was established;secondly,the advantage of dictionary learning for image sparse characterization and spatial constraint for image edge texture fidelity was utilized;then the alternating minimization idea was used to optimize the objective function to achieve the high accuracy solution of the large sick sparse matrix;finally,an external field test was conducted.The experimental results showed that this method had higher reconstruction accuracy compared with the algebraic reconstruction algorithm(ART),joint algebraic reconstruction algorithm(SART),EM algorithm,and the full-region reconstruction error reached 15.81%,which had high application value in the process of spatio-temporal field reconstruction.

关 键 词：三维重建 空间约束 走时层析 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]