LiDAR-Camera Fusion: Guided Vehicle-to-Aircraft Sensing with Cooperative Vehicle Infrastructure

Many airplanes need a guidance vehicle to direct them to specific parking spaces after landing at the airport. In order to enhance airport intelligence and safeguard airport safety, this task requires a high degree of precision in the guidance vehicle's perception of the aircraft. However, in the current research, there are fewer applications of vehicle-road cooperative sensing in airport scenarios, which makes it difficult to exploit the potential roadside sensing capability. In this work, we propose a vehicle-road cooperative perception method, called VICOD, which completes the decision-level fusion and outputs the fused detection frames under the assumption of spatio-temporal alignment. It is based on the vehicle-side detection outcomes generated by the improved AVOD algorithm and the road-side identification results generated by the IA-SSD algorithm. We use the KITTI dataset to validate the associated method. The results of the studies demonstrate that the technique can successfully increase the guided vehicle's sensing accuracy and range to the back of the aircraft.