In-situ precision wear characterization using digital holographic microscopy and computational image processing techniques

Purpose Facilitating a paradigm shift in industrial maintenance from reactive run-to-fail tactics to a proactive, physics-based understanding of surface deterioration is the main goal of this research. This study aims to create a roadmap for the next generation of intelligent tribosystems by bridging the gap between advanced laboratory-grade imaging and the demanding needs of the factory floor. By substituting high-fidelity physical data for speculation, this shift aims to offer a deeper understanding of the basic nature of wear. Design/methodology/approach The deliberate combination of two different sensing technologies forms the methodological foundation of this system. It synchronizes the 3D micro-topography data obtained by digital holographic microscopy (DHM) with high-frequency acoustic emission stress waves. To mathematically refine and quantify material displacement, edge-detection algorithms are applied to this dual-sensor data. This synchronized data is fed into machine learning models, which are continuously trained to differentiate between minute deviations that indicate the beginning of wear and healthy operational signatures to move toward automation. Findings Certain wear mechanisms, like adhesive scuffing and abrasive gouging, which are frequently indistinguishable when using a single-sensor approach, can be isolated by combining acoustic and holographic data. Additionally, compared to traditional vibration analysis, the results show that machine learning models can considerably reduce the rate of false positives by using this high-fidelity dual-stream data. Before they become catastrophic failures, the system effectively records burst events, like microscopic crack formation, in real time. Research limitations/implications This research may make conservative, time-based maintenance estimates obsolete, which is a significant implication. According to the study, actual physical conditions rather than arbitrary schedules will determine the service life of expensive components in the future. Although research emphasizes the successful transition from the lab to harsh realities, it suggests that the next frontier for sustainable manufacturing research is the scalability of such insitu monitoring. Practical implications The practical application of this framework offers substantial benefits for industrial sustainability and cost management. By enabling operators to optimize lubrication cycles and extend the service life of expensive machinery, the system directly reduces waste and avoids costly downtime. The scalability of this insitu monitoring provides a viable path for factories to implement autonomous oversight, ensuring that maintenance is only performed when physically necessary, thereby maximizing both resource efficiency and mechanical reliability. Originality/value The creation of a comprehensive wear fingerprint by synchronizing temporal and spatial data is what makes this work unique. This framework uses the high-speed temporal resolution of acoustic signals in conjunction with the accuracy of DHM to provide a microscopic view of surface health, whereas conventional methods concentrate on macro-level vibrations. In the context of Industry 4.0, this constitutes a novel foundation for autonomous quality control and high-fidelity digital twins. Peer review The peer review history for this article is available at: https://publons.com/publon/10.1108/ILT-11-2025-0500/