Comparative analysis of multimodal feature-based 3D point cloud stitching techniques for aeronautic applications

Designing shimming parts in the aircraft manufacturing process is a common task that needs to be solved. In fact, whenever two or more composite parts need to be coupled, a shim could be designed to absorb any assembly tolerance. This process can be effectively improved with an automated analysis of multimodal data (2D and 3D) directly acquired on the field during an assembly process. In this paper, a comparative analysis of SIFT and SURF feature detectors for the automatic registration and subsequent stitching of dense 3D point clouds in the aeronautic field are presented. Results on real data acquired on the field demonstrate the feasibility of the proposed approach and can be seen as a first step towards the automation of the shimming process.