Online prediction method for chemical quality of sintered ore

The invention is suitable for the technical field of quality prediction, and provides a sinter chemical quality on-line prediction method, which comprises the following steps: determining a plurality of sintering process parameters related to the sinter chemical quality in the sintering process, and obtaining the time sequence characteristics of the plurality of sintering process parameters at the current prediction moment, obtaining sintering machine tail image data from a previous prediction moment to a current prediction moment, determining a key frame sequence of the sintering machine tail image data by using a Unet segmentation network used for extracting a combustion zone form of the sintering machine tail image, extracting a first-level feature of each key frame in the key frame sequence, and extracting a second-level feature of each key frame in the key frame sequence; and obtaining secondary characteristics of the sintering machine tail key frame at the current prediction moment, and inputting the ti