Computer Vision
We develop solutions based on computer vision and deep learning for any industry. We accompany you throughout the process, designing a tailored solution.
Objective. Inspection and detection of defects in automotive press systems using vision
The INSPEC project, “Inspection and Detection of Cracks and Defects in Metal Parts Manufactured by Pressing Processes in the Automotive Sector using Neural Networks and Computer Vision,” aims to develop a vision-based system capable of identifying defects in metal parts produced during the pressing and forming processes. This system seeks to reduce the generation of waste and the need for rework on parts.
This project has received public funding under the PAIP 2019 call from the Government of Aragón to support the growth and consolidation of SMEs, improving their access to financing, technology, and advanced support services, as well as through the FEDER Aragón program.
Solution. We developed a vision system with a custom machine learning algorithm
The solution we developed is based on computer vision and machine learning algorithms. First, we studied the 10 selected parts to be inspected. Understanding them in detail is crucial for selecting the most suitable hardware. The next step was to examine the appropriate vision cameras and lighting to capture data for each process, part, and material of interest.
An essential part of the process is defining what qualifies as a defect and what does not, as there is significant variability.
We used real process data to develop the automatic defect identification algorithms based on neural networks. Surface defects are a significant issue as they generate scrap and require reworking parts, making the process less efficient. With the developed algorithms, we achieved a detection rate of over 90% for the relevant defects.
