WP3: Robust reference dense-image matching, classification and evaluation algorithms
This work package aims to investigate and make available robust reference methods for the whole trustworthy processing of dense image data. The processing is one pillar of the full metrology chain that includes 3 main components: (1) T-MVS, (2) reference thermo-invariant material standards and (3) reference software and reference data.
The addressed and implemented reference processing tools will be validated on numerous reference data to ensure 3D reconstruction, segmentation, classification, evaluation, etc., with very low uncertainties (calculation error <1 nm), aiming at at least a tenth below the achieved measurement uncertainty.
- Task 3.1 will develop a new generator of reference data for validating reference, commercial and open-source dense-image matching software. This will include the evaluation of completeness (how much of the scene has been reconstructed with respect to the ground truth) and accuracy (how close the result is to the ground truth).
- Task 3.2 will develop robust reference dense-image matching algorithms for the MVS systems that are used in industrial applications (e.g. automotive, aeronautic, etc).
- Task 3.3 will investigate and implement robust algorithms for the automation of defect classification for the acceptance/rejection criteria made using MVS.
- Task 3.4 will investigate and develop advanced reference algorithms for the surface quality evaluation of functional requirements, where the quality control of geometric characteristics and the conformance of manufactured parts to design tolerance specifications should be 100 % inspected in real-time with a well-known uncertainty by applying reference algorithms.
WP leader: Prof. Hichem NOUIRA (LNE) – Hichem.Nouira@lne.fr