Photon counting 3D imaging from multiple positions

Detecting and identifying objects inside a forest edge on the other side of an open field is an important task in defence and security applications. This can be difficult to achieve with passive imaging sensors because of the partial obscuration by the foliage. High-resolution 3D imaging enables separation of hidden objects from branches and leaves and can provide data for detection, recognition, and identification of partly occluded targets. We use a photon counting lidar system with panoramic scanning to produce high quality 3D data for this task. The FOI system is built around the Princeton Lightwave Inc. (PLI) Falcon detector which is a 32×128 pixel array of InGaAs Geiger-mode avalanche photodiodes. The system operates at 1557 nm and has been designed for suitable resolution at standoff ranges of 1 to 2 km. In this paper, we investigate the detection capability when combining measurements from multiple measurement positions. A field trial has been performed where data from the same scene was collected from different sensor positions. The system was mounted on a car and moved between different positions along a road. The measurements were performed first without and then with vehicles in the target area. The distance to the forest edge varied between from approximately 1.0 to 1.5 km, and the difference of the angle of incidence was approximately 45 degrees from the outer positions along the road. To merge the data from the different positions we apply registration of the data sets using derived point clouds to transform all data into a common coordinate system. Data from the different sensor positions is analyzed by overlaying the derived point clouds from the different positions. We compare data from different viewpoints to data from only one viewpoint. The results show that the combined point clouds from multiples positions covers more of scene than from a single position. We also perform change detection using registered point clouds from the same measurement positions. In the change detection we found the changes we had introduced (vehicles and equipment).