Prototype System for Visual Inspection in Serial Production of Polymer Components Using Artificial Intelligence Methods

OP TAK No. CZ.01.01.01/01/22_002/0000590 Prototype System for Visual Inspection in Serial Production of Polymer Components Using Artificial Intelligence Methods (01/2024 – 06/2026), Ministry of Industry and Trade

Consortium for Industrial Research and Development of New Laser Technology Applications Using Artificial Intelligence Methods

Operational Programme Enterprise and Innovation for Competitiveness

Registration number: CZ.01.1.02/0.0/0.0/17_147/0020575
Project leader: MEPAC CZ, s.r.o.
Partner: University of Ostrava

Financial Grant:

Total grant amount: CZK 24,215,211.00
EU contribution: CZK 16,950,647.70
Own funding: CZK 7,264,563.30

Project Dates:

Project start date: October 1, 2020
Project end date: June 30, 2023

Provider:

Ministry of Industry and Trade

Contacts:

Project main leader: MEPAC CZ, s.r.o. – Ing. Milan Kaštan, Ph.D.
Project co-leader: University of Ostrava – prof. RNDr. Jiří Močkoř, DrSc.

Project Description:

The project is planned for a period of 3 years, and the output will be two prototypes of laser systems utilizing artificial intelligence elements to enhance the quality and precision of operations. Existing laser systems produced by MEPAC CZ will thus be upgraded to 6th-generation prototypes with high commercial potential.

This goal will be achieved through the long-term theoretical and practical experience of the proposer (MEPAC CZ) in precision engineering and laser technologies, combined with theoretical and pre-application results obtained by the project partner (UVAFM) in the AI-Met4AI project, particularly in the area of image processing, including theoretical methods based on AI applications for rapid and reliable object identification in image datasets.

Project Goal:

The aim of the project is to create two completely new functional prototypes of laser systems: one for laser engraving and one for laser cladding. The uniqueness of these systems will lie in the inclusion of artificial intelligence elements, significantly enhancing device functionality and user-friendliness. The AI will be implemented in a unique way, based on deep neural networks as well as unique theoretical approaches such as fuzzy transformations and other methods for simplifying analyzed images.

Project Activities:

The project is divided into four stages, i.e., logical units (DAP = partial project activities). These DAPs run concurrently, each focusing on a specific aspect of the laser system. Key project milestones span across the DAPs and represent critical components of the laser system.

The following DAP activities will be carried out:

DAP 1 – Laser Systems Hardware

• Concept and design of laser system components based on conducted research.
• Linkage with other DAPs of the project.
• Complete mechanical design of prototypes.
• Preparation of all hardware components for engraving and welding laser systems.
• Assembly of partial hardware units.
• Testing of partial hardware units, laboratory application.
• Application in real operational conditions.
• Evaluation of tests, proposal of suitable modifications.
• Preparation for integration.

DAP 2 – Laser Systems Control Software and Interface

• Concept and selection of components based on conducted research. Linkage with other DAPs.
• Preparation of all software components for engraving and welding systems (excluding AI elements – separate DAP).
• Complete software design of prototypes.
• Activation of partial hardware elements of prototypes.
• Testing of partial elements of laser systems.
• Application in laboratory conditions.
• Application in real operational conditions.
• Evaluation of tests, proposal of suitable modifications.
• Preparation for integration.

DAP 3 – Artificial Intelligence Elements (Intelligent Software)

• Design of surface marking methods for capture, digitization, and depth reconstruction.
• Smooth surface modeling from incomplete data and detection of semantically significant areas.
• Implementation of machine learning algorithms for robotic arm path planning.
• Creation of software simulator and robot motion training.
• Integration and testing of robot motion in real conditions.
• Design and implementation of classifier.
• Creation of synthetic dataset for weld quality inspection.
• Training the system for weld quality classification on synthetic and real datasets.
• Integration of quality control into the robotic system.

DAP 4 – Integration of Laser System Components

• Concept of two laser system prototypes based on conducted research.
• Production and commissioning of two laser system prototypes.
• Integration of partial laser system components (hardware and software).
• Gradual verification of functionality of communication bridges.
• Integration of AI elements.
• Commissioning of welding laser system prototypes.
• Commissioning of engraving laser system prototypes.
• Testing in laboratory conditions, evaluation, and proposal of modifications.
• Testing and application in real operational conditions, evaluation, and proposal of modifications.
• Finalization of prototypes including technical documentation for market release.

Within individual DAPs, both experimental AI development issues (mainly handled by the project partner) and industrial development issues related to technological implementation (mainly handled by the project leader) will be addressed in a coordinated manner.

This project is financially supported by the EU.

Published / Updated: 02. 02. 2021