Abstract: A tractor has a processor and memory storing machine-readable instructions that, when executed by the processor, cause the processor to: select at least one image captured by a camera of a parking spot being passed by the tractor when a point-of-interest (POI) of the parking spot is most centered within a field of view of the camera, store the image in the memory in association with the parking spot, determine that the parking spot is occupied after a costmap for the parking spot is populated by the LIDAR, retrieve the image corresponding to the parking spot from the memory, determine, from the image, a trailer identifier of a trailer parked in the parking spot; and send a message defining the parking spot and the trailer identifier to an inventory service for use in updating a yard inventory.