The invention relates to the technical field of carbon detection, in particular to an online real-time dynamic detection method for the carbon content of converter
steelmaking molten steel, which comprises the following steps: acquiring multi-source
sensing data, and acquiring a
molten steel surface temperature image acquired by a
thermal infrared imager; acquiring multiband
spectral data, namely controlling a
visible band sensor to acquire
radiation intensity in a range of 500-600nm, and controlling a near-
infrared band sensor to acquire
reflectivity characteristics in a range of 850-1100nm; analyzing and fusing the data, analyzing the temperature image, and generating a
molten pool two-dimensional temperature field; analyzing the multiband
spectral data, and extracting a spectral
feature vector related to the carbon content; fusing the two-dimensional temperature field and the spectral
feature vector, and generating a
molten steel surface carbon
content distribution diagram based on a pre-trained carbon content
machine learning model; and outputting an optimized detection result,
processing multi-node data through a
distributed computing framework, and outputting a dynamic carbon content detection report. According to the invention, the effects of breaking through the limitation of traditional static detection, real-time component monitoring, full
molten pool covering and dynamic feedback are achieved.