Research on Machine Learning Techniques Applied to Industrial Manufacturing
Summary
TALENT project is aimed at the industrial sector to improve quality assurance processes for manufactured products through the evolution and enhancement of a 3D industrial inspection system.
Description
The project includes the following tasks:
-
Research and evolution of geometric comparison analysis to increase system accuracy and expand its applicability to parts that can tolerate certain degrees of deformation (such as the appearance of synthetic bulges on flat surfaces during 3D reconstruction) that do not imply a quality failure.
-
Advancement in the creation of a tool that allows for the comfortable and easy generation of GD&T specifications for the dimensions and tolerances accepted, based on a 3D surface of the reference model.
-
Research and development to improve measurement accuracy by enabling the use of reconstructions through multiple launches in the technology, which enhances the precision of measurements taken by the system.
-
Research and development for improving the calibration and focusing process to reduce times.
-
Enable dynamic management of information displayed in relation to the parameters of interest for the manufactured parts, allowing them to be compared with expected KPIs. For this, a graph has been implemented in the UI showing the evolution of measurements taken for each piece in a batch, so that all or part of the information for that batch is displayed.
-
Redesign and update the available hardware prototype in the laboratory to achieve a version that can process large test batches in a manner similar to a real environment.