Design and Application of Tobacco Impurity Removal Model Based on Convolutional Neural Network

This paper explores the design and application of the tobacco impurity removal multi-layer convolution model based on the convolution neural network of the tobacco data set. Traditional tobacco impurity removal methods are inefficient and costly, so this paper proposes a new method. First, we introduce tobacco impurity removal technology's importance and current status and discuss the basic principles and applications of convolutional neural networks based on tobacco datasets. In addition, we constructed a tobacco image dataset in this study and created a multi-layer convolutional model to remove tobacco contaminants. It has high accuracy and robustness and realizes automatic discrimination and separation of tobacco and impurities.