Infrared temperature measurement technology is widely used in the detection of surface temperature distribution of hot stamping parts. However, the accurate infrared temperature measurement for complex parts is still a problem due to the deviation of emissivity. In this study, a novel directional emissivity calibration was proposed for 22MnB5 hot stamping steel. A mathematical model considering directional emissivity was established and verified. Based on the projection and interpolation method, the directional emissivity is assigned to each pixel in the thermograph shot by infrared thermal imager. Finally, a high measurement precision within ± 25 °C was achieved for temperature measurement on 22MnB5 hot stamping parts.

Infrared thermal imaging system is a key component in hot stamping and is commonly used to collect parts temperature during the part production process. However, due to the inappropriate setting of the emissivity, the output produced via such system usually lack accuracy and consistency. A ± 150 °C temperature measurement error was observed at several hot-stamping lines. In this paper, a new infrared software system is proposed to calculate parts temperature by utilizing a mathematical model behind the scenes. The system factorizes 3D model of the part, position of the thermal camera and parameters of the environment to calculating the directional emissivity, which is further used to compensate the output obtained from the thermal camera. Compared to traditional thermal systems, the proposed software system achieved 83% increase in accuracy (±25 °C vs ±150 °C in measurement error). Nonetheless, usage of such software system will be discussed. For example, double sheet detection for hot stamping line have been implemented by setting the threshold of part temperature and dispatch warnings to PLCs via ProfiNet or Ethercat protocol. In combined with IIot and machine learning techniques, quality prediction system for hot stamping have also been built.