Temperature detection in ironmaking and steelmaking processes is crucial for enhancing quality control, achieving precise closed‐loop control, optimizing production procedures, and reducing carbon emissions. Traditional infrared temperature detection technologies suffer from limitations such as insufficient measurement accuracy, difficulty in separating background interference, and significant response delays, making them inadequate for meeting the stringent requirements of high‐temperature, high‐pressure, and high‐precision real‐time temperature measurement in the steel industry. This article analyzes the current state of online temperature detection technologies in ironmaking and steelmaking processes and examines the temperature measurement requirements across different production stages. Then, a high‐temperature online detection technology system based on multispectral and multi‐information fusion is proposed, incorporating multispectral sensors, noise filtering, image recognition of interference zones, and colorimetric thermometry models. The application of multispectral temperature measurement is demonstrated in key areas including online temperature monitoring of the blast furnace tuyere zone, inside refining furnaces, on the surface of continuous cast slabs, and during the cooling process of wire rods. The proposed online temperature detection technology fuses multispectral radiation data, image‐derived features (e.g., texture and morphology of slag/scale), and process parameters. Compared with prior studies, the main novelties of this work lie in: (1) the real‐time integration of multispectral imaging and deep learning‐based semantic segmentation within a unified system, enabling automatic interference isolation and net‐surface temperature extraction; and (2) the field validation of the same technological framework across multiple key stages of steel production—from ironmaking and steelmaking to continuous casting and rolling—demonstrating its broad adaptability and engineering practicality. This integrated approach is expected to advance the development and application of AI‐enhanced high‐temperature online detection technologies in steel production processes. This article serves as a state‐of‐the‐art review, incorporating representative case studies from the authors’ own research group to illustrate the practical implementation and effectiveness of the proposed technology system.
Shi et al. (Sat,) studied this question.