Through Thick and Thin: Imaging through Obscurant using SPAD array

Abstract

Preliminary work on 3D image collection and classification of targets in the presence of obscurant using a Flash LiDAR system is discussed in this paper. The system is based around a DST designed 32 x 32 Single Photon Avalanche Diode (SPAD) array to image either targets or silhouettes of targets. The collected data included military targets that were obscured either by camouflage nets or fog. For camouflage net, the target was detected using an algorithm implemented on the Nvidia Jetson TX2. Targets obscured by fog are detected and classified where the classification accuracy is 100% for fog visibility down to 17.3m and 89.5% for 14.1m. This algorithm was not implemented on the TX2 but its simplicity shows potential for it in the future. This initial approach opens the road to eventually operate SPAD based systems for real-time classification through dust or smoke.