Benefiting from the high sensitivity and high temporal resolution of single photon detector, Geiger mode laser ranging technology has become a research hotspot for three dimensional imaging detection in recent years. The range reconstruction method is one of the key technologies in the application of laser three dimensional imaging. Facing the application of long distance, high precision laser three dimensional imaging,the biggest technical problem based on photon counting LiDAR is to maintain the distance reconstruction accuracy under the interference of strong background noise. Most of the noise of single photon detectors can be suppressed through distance gating and spectral filtering. The laser backscattered noise also can be suppressed through polarization modulation. But,there still exist strong backgrounds such as sunlight scattering within the distance gate and the spectral bandwidth. Laser ranging accuracy is decreased by sunlight scattering noise. The results of previous photon counting laser ranging experiments under daytime conditions show that,it is impossible to meet the needs of high precision ranging under strong noise backgrounds by improving the performance of single photon detectors of lasers. Therefore,it is necessary to study different distance reconstruction methods,and propose a laser echo information processing method under strong noise conditions. The experimental comparison study work can provide the support for the innovation and development of long-distance, high frame rate active laser imaging technology. In this paper,the principles and characteristics of multi-class distance reconstruction methods based on cell Geiger mode avalanche photodiodes are systematically sorted out in. It mainly includes photon counting time statistics method,linear or exponential fitting method,matrix or Gaussian matching method,temporal correlation method and spatial correlation method. Aiming to the application of long distance and high frame rate lase three dimensional imaging,an experimental scheme based on laser beam emission scanning and focal area array detector receiving is proposed. The scanning azimuth information is fused in the process of laser echo data processing to improve the detection signal to noise ratio. A laser three dimensional imaging experimental device is built. The two dimensional scanning angle of the laser beam covers the detection field of view. Based on the digitized output of Geiger mode focal plane array avalanche photodiode,a comparative experiment of different distance reconstruction methods is carried out. The different distance reconstruction methods are used to extract target distance information. The difference in target distance reconstruction results is quantitatively evaluated by the ratio of signal photon counts to background photon counts. The experimental comparison results show that,for focal plane array detector,the spatial correlation method has better distance reconstruction effect. Meanwhile, the spatial correlation method can greatly improve the frame rate of three dimensional imaging. The better reconstruction effect is mainly reflected in two aspects. Firstly, the ratio of signal photon counts to background photon counts after spatial correlation processing has an order of magnitude improvement,from 0.5 to 6.67,while the improvement effects of other reconstruction methods in this paper are not obvious. Secondly,it is found that only the spatial correlation method can reconstruct the three dimensional contour of the target at an imaging frame rate of 100 frames per second. The experimental results confirmthe advantages and effectiveness of the spatial correlation reconstruction method. Through theoretical and experimental analysis,the research results in here provide a theoretical basis and experimental support for the application of the high-frame rate detection of laser three dimensional imaging under strong noise background. The next step is to carry out outdoor long- range laser 3D imaging experiments and to further explore and study the application of three dimensional imaging target recognition and tracking method based on array devices.
