基于SPAD阵列的共光路扫描三维成像

Coaxial Scanning Three-dimensional Imaging Based on SPAD Array

基于单光子雪崩二极管(SPAD)阵列探测器的三维成像技术在工业和科学领域具有重要的应用前景。然而现有的SPAD阵列器件仍存在小阵列规模和低像素填充率所导致的低空间分辨率问题,为此基于SPAD阵列(32pixel×32pixel)和衍射光学元件(DOE)搭建收发共光路扫描三维成像实验系统。利用DOE将出射激光整形为激光点阵并与接收视场匹配可以提升激光能量利用效率,通过共光路扫描可以实现高分辨三维成像。利用基于滑动时间窗的噪声光子滤除算法和基于全变分正则化的图像重构算法对回波光子数据进行处理。实验结果表明,图像重构算法可以获取10 m外目标64 pixel×64 pixel的三维图像,且在平均每像素0.86个信号光子的条件下实现清晰成像,成像结果的平均绝对误差为0.016m。

Threedimensional(3D)imaging technology based on single-photon avalanche diode(SPAD)array detectors has important industrial and scientific applications.However,existing SPAD array devices are limited by low spatial resolution due to small array size and low pixel-filling factor.Therefore,a coaxial scanning 3D imaging experiment system is built using a SPAD array(32 pixel×32 pixel)and a diffractive optical element(DOE).Using the DOE to shape an outgoing laser into a laser lattice and match it with the receiving field of view can improve laser energy utilization efficiency and achieve high-resolution 3D imaging through coaxial scanning.A noise photon-filtering algorithm based on a sliding time window and an image reconstruction algorithm based on total variational regularization are used to process the echo photon data.Experimental results show that,a 64 pixel×64 pixel 3D image of a target at a distance of 10 m can be obtained with an average of 0.86 signal photons per pixel to achieve clear imaging,the average absolute error of imaging results is 0.016 m.