Convolutional neural network coke thermal state quality prediction method based on coal data imaging

The invention relates to a convolutional neural network coke thermal state quality prediction method based on coal data imaging. The method is composed of coal data cross-domain conversion and convolutional neural network prediction. Through a data dimension reduction compression method, the mixed coal data is converted from a structured numerical field to an unstructured image field, so that the characteristic relationship among mixed coal parameters is more easily captured and analyzed by a model; and based on the blended coal data of the image domain, realizing prediction of coke reactivity CRI and coke post-reaction strength CSR indexes by using a trained convolutional neural network coke thermal state quality prediction model composed of a convolutional layer, a pooling layer and a full-connection layer. According to the method, the nonlinear mapping relation between the blended coal parameters and the corresponding coke thermal state quality indexes can be accurately analyzed, the network reasoning proc