The ID2M+ project, titled “InnoProfile‑Transfer – Intelligent Digital Multichannel Image Processing and Capture,” extends the original ID2M research by integrating advanced camera evaluation, expanded spectral imaging, rapid 3D acquisition, and online process monitoring. The initiative was carried out in the fourth year of the project, achieving all planned milestones and delivering significant performance gains across multiple imaging domains.

Technical Results
Key technical achievements of the ID2M+ investment packages are summarized below:

Camera Evaluation (AP1) – Comprehensive characterization of high‑precision cameras over a broad spectral range of 400–2400 nm, extending visible (VIS) coverage into the short‑wave infrared (SWIR). Temporal metrics such as jitter and latency uncertainty were introduced, providing a robust framework for sensor optimization.
Extended Spectral Processing (AP3) – Infrared sampling was broadened to 1700–13000 nm, enabling fundamental studies in the mid‑ to far‑infrared. Online spectral data processing was accelerated on heterogeneous multiprocessor systems, achieving algorithm‑dependent latencies suitable for real‑time applications. UV spectroscopy (200–400 nm) was supported by power‑regulated illumination modules.
Deep Learning Platforms – The Nvidia DGX‑1 GPU cluster and Xilinx UltraScale+ FPGA technology were benchmarked for deep‑learning inference. These platforms facilitated rapid training and deployment of convolutional neural networks for image‑based decision support.
Fast 3D Imaging (AP4) – Embedded 3D imagers were upgraded to a data rate of 25 B/s at 1.3 MP resolution, with full topography computation performed on UltraScale+ FPGAs. High‑speed sequential 3D capture and multi‑view sensor networks were established, enabling simultaneous multi‑angle acquisition.
3D Online Process Monitoring (AP5) – Real‑time control of additive manufacturing processes was implemented, integrating the 3D imaging pipeline into FDM/SLM machines. Adaptive path control leveraged low‑latency 3D data to adjust tool trajectories on the fly, improving part quality and reducing defects.
Application Laboratory – A combined lab environment was created, allowing students and researchers to conduct pre‑studies and training across all imaging modalities, fostering interdisciplinary skill development.

Collaboration
The project’s success hinged on close cooperation between the Qualimess research group, local industry partners, and national technology providers. Procurement followed a competitive bidding process, with at least three offers evaluated for each item and public tenders for purchases above €20,000. The integration of MWIR and LWIR cameras enabled weld inspection through slag and real‑time monitoring of melt pool emissions, providing actionable insights into weld quality. The DGX‑1 system was deployed in partnership with Nvidia, while the UltraScale+ FPGA solutions were sourced from Xilinx. Industrial collaborators contributed real‑world datasets and validation scenarios, ensuring that the developed algorithms and hardware met production requirements. This collaborative framework not only accelerated technology transfer but also attracted graduate students to the field of industrial image processing, strengthening the region’s skilled workforce.