Tobacco leaf primary baking control method based on image reasoning model

The invention provides a tobacco leaf primary baking control method based on an image reasoning model, which comprises the following steps of: establishing a yellow proportion and dry proportion prediction model, writing the prediction model into software, and installing the software in an intelligent control system of a bulk curing barn; the actual yellow proportion and dry proportion in the current baking process are obtained through an intelligent control system with a field image collecting function and compared with the yellow proportion and dry proportion required by a baking curve, if the yellow proportion and the dry proportion are the same, the target is achieved in advance, and if the yellow proportion and the dry proportion are different, the target is achieved in advance. If yes, estimating delayed fire time, automatically delaying fire, adjusting dry-bulb temperature and wet-bulb temperature and continuing to bake according to yellow baking characteristics and dry baking characteristics of the to