This PhD project focuses on developing advanced non-contact measurement systems based on infrared vision, applied to Industry 4.0 and forensic science. The primary goal is to provide innovative methodologies for intelligent process control and the reliable identification of physical evidence. In the industrial sector, research was conducted within the openZDM project in collaboration with VDL Weweler. Two non-destructive inspection systems based on passive thermography with Near-Infrared CMOS sensors were designed, developed, and integrated in-line. These systems monitor 100% of production. A key contribution involves metrological calibration and the implementation of self-diagnostic procedures to ensure data reliability and operational continuity. From a methodological standpoint, original approaches were developed to reconstruct 2D thermal distributions from image sequences acquired during rapid thermal transients (e.g., cooling after induction heating). Two methodologies, based on temporal and spatial alignment, allow for complete thermal imaging without artifacts. Specific algorithms were validated to detect surface anomalies, such as thermal non-uniformity and residual oxide scales, supporting the Zero-Defect Manufacturing paradigm. In the forensic field, the thesis explores infrared vision for detecting and analyzing gunshot residue (GSR). A novel methodology based on active thermography with flash excitation is presented. This technique highlights emissivity changes between the fabric and combusted powder, enabling the estimation of shooting distances. Finally, SWIR multispectral analysis was employed to overcome the limitations of visible-light inspection on dark fabrics, optimizing contrast for the quantification of unburned particles.
Questo progetto di dottorato riguarda lo sviluppo di sistemi di misura non a contatto basati sulla visione a infrarossi, con applicazioni in ambito Industria 4.0 e forense. L’obiettivo principale è fornire metodologie innovative per il controllo intelligente dei processi e l'identificazione di prove fisiche. In ambito industriale, la ricerca si è svolta nel progetto openZDM con il partner VDL Weweler. Sono stati progettati e integrati in linea due sistemi di ispezione non distruttiva basati su termografia passiva con sensori CMOS Near-Infrared. I sistemi operano sul 100% della produzione. Un contributo significativo riguarda la calibrazione metrologica e l’implementazione di procedure di auto-diagnosi per garantire affidabilità e continuità operativa. Dal punto di vista metodologico, sono stati sviluppati approcci originali per la ricostruzione di distribuzioni termiche 2D da sequenze acquisite durante transitori termici rapidi (es. raffreddamento post-forno a induzione). Due metodologie, basate sull'allineamento temporale e spaziale, permettono di ottenere immagini termiche prive di artefatti. Algoritmi specifici sono stati validati per rilevare anomalie superficiali, come difetti di uniformità e residui di scaglia di ossido, contribuendo concretamente al paradigma Zero-Defect Manufacturing. In ambito forense, la tesi esplora l'uso dell'infrarosso per l'analisi dei residui di sparo (GSR). È presentata una metodologia basata sulla termografia attiva con eccitazione flash, capace di evidenziare variazioni di emissività tra tessuto e polvere combusta, permettendo la stima della distanza di sparo. Infine, l'analisi multispettrale SWIR è stata impiegata per superare i limiti della visione nel visibile su tessuti scuri, ottimizzando il contrasto per la quantificazione delle particelle incombuste.
Exploiting Infrared Vision Potential: from High-Temperature Manufacturing to Forensic Trace Detection / Medici, Vittoria. - (2026 Mar 20).
Exploiting Infrared Vision Potential: from High-Temperature Manufacturing to Forensic Trace Detection
MEDICI, VITTORIA
2026-03-20
Abstract
This PhD project focuses on developing advanced non-contact measurement systems based on infrared vision, applied to Industry 4.0 and forensic science. The primary goal is to provide innovative methodologies for intelligent process control and the reliable identification of physical evidence. In the industrial sector, research was conducted within the openZDM project in collaboration with VDL Weweler. Two non-destructive inspection systems based on passive thermography with Near-Infrared CMOS sensors were designed, developed, and integrated in-line. These systems monitor 100% of production. A key contribution involves metrological calibration and the implementation of self-diagnostic procedures to ensure data reliability and operational continuity. From a methodological standpoint, original approaches were developed to reconstruct 2D thermal distributions from image sequences acquired during rapid thermal transients (e.g., cooling after induction heating). Two methodologies, based on temporal and spatial alignment, allow for complete thermal imaging without artifacts. Specific algorithms were validated to detect surface anomalies, such as thermal non-uniformity and residual oxide scales, supporting the Zero-Defect Manufacturing paradigm. In the forensic field, the thesis explores infrared vision for detecting and analyzing gunshot residue (GSR). A novel methodology based on active thermography with flash excitation is presented. This technique highlights emissivity changes between the fabric and combusted powder, enabling the estimation of shooting distances. Finally, SWIR multispectral analysis was employed to overcome the limitations of visible-light inspection on dark fabrics, optimizing contrast for the quantification of unburned particles.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.


