In order to study the application of digital image processing technology in chemical composition prediction, we have verified the detection of nicotine content in different stages of tobacco. Taking the tobacco leaves at different stages in the baking process of flue-cured tobacco as the research object, the image processing technology was used to extract the image parameters of tobacco leaves at different stages, and the BP neural network was used to establish the prediction model to achieve non-destructive testing. The final prediction results show that the model is feasible to predict the nicotine index at different stages of the baking process according to the color eigenvalues and texture eigenvalues of the tobacco leaves. The prediction results are in line with the actual situation, and the model has certain practical value. This confirms the application value of digital image processing technology.
