Advances in the Application of Machine Vision in Steelmaking Processes

To enhance steelmaking efficiency, guarantee product quality, and ensure process stability, the global steel industry is rapidly advancing toward intelligent transformation. As a key component of artificial intelligence, machine vision technology is being increasingly integrated into all stages of the steelmaking process. This article systematically reviews advancements in hardware architecture and the evolution of algorithmic models from traditional image processing and machine learning to deep learning in machine vision technology. Subsequently, it summarizes the current applications of this technology across the hot metal pretreatment, converter steelmaking, secondary refining, continuous casting, and others. Finally, this review consolidates the core advantages of machine vision in steelmaking: optimizing smelting processes, enabling precise prediction of abnormal operating conditions, facilitating accurate product classification/grading, and enhancing automation by reducing manual intervention. Future development trajectories are summarized as follows: focus on transitioning from isolated applications to systematic integration, ensuring input data quality and reliability, enhancing model interpretability, and integrating machine vision with digital twin technology.