基于参数化盲反卷积的三维非扫描激光雷达距离估计

Three Dimensional Flash Imaging Ladar Range Estimation Via Parameterized Blind Deconvolution

三维(3D)非扫描激光雷达具有多表面目标距离分辨能力,可以用于对隐藏和伪装目标的识别。为了快速、准确地估计3D非扫描激光雷达多表面目标距离信息,提出了基于期望值最大化(EM)的单像素多表面目标的距离估计算法,通过对系统点扩展函数的参数化,该算法可以同时估计出成像系统点扩展函数和目标的距离信息。仿真实验结果表明,相比于传统的混合高斯匹配算法和维纳空间滤波算法,该算法在系统点扩展函数未知的条件下,可以将目标的距离估计精度分别提升大约70%和40%。

Multi-surface ranging with the use of three dimensional （3D） flash ladar can be useful in accurately discriminating camouflaged targets of interest. In order to estimate multiple surfaces in 3D flash imaging ladar well and truly, a multiple surfaces＇ range estimating algorithm in 3D flash imaging ladar via expectation maximization （EM） is proposed. This algorithm can estimate the point spread function of imaging ladar and the target range information simultaneously. Simulation results show that the multiple surfaces＇ range estimating algorithm can improve range estimation over traditional mixed Gaussian matching and Wiener filter by up to 70% and 40% respectively.