基于盖革APD阵列的光子计数三维成像

Photon-counting 3D imaging based on Geiger-mode APD array

  • 摘要: 基于盖革APD阵列的激光主动探测系统具有较高的灵敏度、空间分辨率和测距精度,在遥感探测、目标识别等领域具有广泛的应用前景。受探测模式、噪声等因素影响,盖革APD阵列需要大量累积光子探测来实现高精度成像。针对该问题,基于目前国内规模最大的InGaAs盖革APD阵列,搭建了1 064 nm激光探测实验装置,对室外600 m外目标进行了成像。通过分析光子计数物理过程,建立了目标反射率与距离的极大似然估计。结合自然图像稀疏的先验知识,采用正则化图像重构方法,改善了累积光子数较少情况下的成像精度。通过对比,验证了正则化图像重构方法能够抑制光子数涨落引起的参数估计偏差,提升了成像质量。

     

    Abstract: Laser detection and ranging system based on Geiger-mode APD (GM-APD) array can achieve very high sensitivity, spatial resolution and ranging precision, which has wide applications in remote sensing, target recognition and so on. Limited by the factors of detection mode and noises, GM-APD usually requires large number of photon detections to carry out high imaging precise. For this problem, firstly a ladar system based on national state-of-the-art InGaAs GM-APD array was set up at a wavelength of 1 064 nm. The system was then used for imaging of a building about 600 m away. The photon detection process on GM-APD was analyzed for the reflection and ranging information extractions with maximum likelihood estimation method. For the condition that only a few photons were detected on a pixel, the prior of natural scene’s spare property was used to improve the imaging precise based on regularized image reconstruction method. Through comparison, it was verified that the regularized method can reduce the estimation error due to photon detection fluctuation, and improve the imaging quality.

     

/

返回文章
返回