Algorithms for Print Quality Monitoring
In collaboration with
Scenario
We have been contacted by Cefla Finishing, a leading company in the production of machines for painting, decorating, and digital printing on various types of materials. The need was to make the process for verifying print quality more efficient.
The project's objective was to develop an algorithm capable of generating a print quality control pattern and analyzing this pattern to provide the control system and the operator with guidance on correcting any detected anomalies.
The challenge
Cefla Finishing's digital printers allow high-resolution printing of images, patterns, and embossed textures directly onto the surface of a wide range of materials. Quickly identifying possible print anomalies due to clogged nozzles or print head alignment issues is essential to ensure the highest quality of the finished product and minimize machine downtime.
Another challenge of the project was the need to manage a wide range of printers with a single software, with significant variability in terms of sizes, print resolutions, and colorations. In particular, the notable physical dimensions of the printing area necessitate the development of ad hoc techniques for the rapid generation and analysis of patterns.
The solution
The collaboration with the Cefla team led to the development of an algorithm tailored to their needs.
Through the use of a high-performance language like Rust and programming and code optimization techniques, the pattern generation and analysis routines minimize execution times.
We developed ad-hoc algorithms for image analysis that are more efficient in terms of precision and execution times.
To allow for easy integration into the client's existing systems, a dedicated API module was developed.
Results achieved
Thanks to the use of specially selected and developed technologies, techniques, and algorithms, execution times have been reduced from several minutes to just a few seconds.
The library is parameterized on the characteristics of the printers, allowing the management of a wide range of printers and configurations with a single platform.
Future developments
The developed algorithms are designed to be extended over time to allow real-time print analysis, eliminating the need to generate predefined patterns and further reducing machine downtime.
If you think you need our support for innovative activities of this type, contact us to discuss your needs together.