International Business Machines Corporation
DELIVERY PACKAGE DAMAGE DETECTION USING PHOTOGRAMMETRY MODELS

Abstract:

In order to detect possible damage to a delivered product, a set of first images of the product is created prior to delivery which are used to generate a first 3D photogrammetry model, and a set of second electronic images are created after delivery which are used to generate a second 3D photogrammetry model. The two models are then compared to determine whether there is a sufficient deviation to conclude the product has been damaged. The comparison is performed by a cognitive system trained with samples of before/after photogrammetry model pairs each associated with either a damaged condition or a not damaged condition. The baseline photogrammetry model is part of a blockchain record and the recipient photogrammetry model is added to the blockchain record. Intermediate photogrammetry models can also be built where there is more than one shipper along the delivery route to assign liability to the proper party.