Packaging optimisation is a crucial factor in the profitability of any logistics business. However, the problem begins when the packages arrive in a constant and difficult to predict manner. Luckily, the Structure sensor combined with the ViewAR technology enables assessing volume of an aircraft pallet of packages only from a live 3D scan with an iPad.
AR meets depth sensing
Depth sensing is a relatively new, dynamically developing branch of technology. Whether it is a time of flight or structured light based solution, in combination with AR it has a great potential. Since the technology may be used for handling occlusion, dealing with collisions, and tracking hand gestures, numerous new use cases might be developed – 3D scanning, measurements (distances, surfaces, volume), surface replacement, just to name a few.
In the Lufthansa Cargo case, the business challange was to create a robust, mobile way to estimate volume of plane pallets. As long as there already was an estimation of their mass, the volume of each package was given separately, with no link to the space a stack of them would take. Conventional stationary systems were reliable but expensive and taking up a lot of space in the warehouse. Furthermore, their use was time-consuming and required heavy equipment for loading and unloading. In a business model with multiple flights leaving an airport every day, time and cost-optimisation play a crucial role.
For the need of this project, the Structure sensor by Occipital combined with an iPad was used in order to obtain a marker-less, real-scale, robust tracking system. The technology was chosen for its precision and robustness of mesh reconstruction.
As for the software, two main functionalities were developed: mobile volume calculation and visualisation, and digital contour check. First, the pallet would be scanned by a warehouse worker with the use of a mobile device. Then, a volumetric mesh representing its geometry would be reconstructed by the Structure Sensor software. Since pallet’s dimensions go up to 3x3x3 meters, a custom ViewAR hole-filling algorithm was developed in order to generate missing surfaces e.g. at the top of the volume, where the worker would not have access. Next, a visualisation of the contour of the aircraft around the pallet would be displayed in the filed of view of the user. It would be utilised to verify whether the load would fit into the particular plane (multiple plane model options available). Such a solution dramatically reduced the time of a volume check of a pallet, shortening it to 90 seconds, in the same time preserving the accuracy.
The described solution was tested in the Frankfurt warehouse of Lufthansa Cargo, operating 15-20 flights per day. The company underlines achieved improvements in planning and is currently investigating a business model with estimations shifted from weight to volume. Alexander Hage, Senior Manager from Lufthansa, sees great potential in AR, especially in the planning (volume optimisation, digital contours check, 3D planning), operations (shipment dimensions, label recognition, information handling) and maintenance (transport, navigation, remote assistance) domains.