The knowledge of the geometry of a product's components throughout its life cycle is essential in order to guarantee its conformance. Geometry acquisition processes are relying increasingly on optical digitizing systems which produce discrete data (point clouds, meshes, images...) representative of the part geometry in short time. Considering the increasing needs in terms of measurement complexity and accuracy, the use of multisensor systems is becoming more and more usual, which leads to new challenges.
The objective of our work is to ensure the representativeness of the discrete data obtained by digitizing, i.e. to ensure both the completeness and the quality of the measured data. Three main topics are more particularly investigated:

• Performance assessment of non-contact digitizing systems
• Definition of multisensor, multiscale digitizing strategies
• Consistency assurance of the acquired data (registration, merging, partitioning)