Application of convolutional neural network in the production and processing of flower and fruit tea

• CNN reduces feature extraction time • CNN outperforms traditional and machine learning algorithms • The application of CNN promotes the transformation of the tea industry towards intelligence • Provide theoretical guidance for the sustainable development of the tea industry At present, the global tea industry is in a stage of transformation towards intelligent chemical development. Although traditional machine learning methods have achieved good results in the production and processing supervision of flower and fruit tea, it is difficult to improve supervision efficiency due to the limitations of manually extracting features. The automatic feature learning function of convolutional neural networks(CNN) solves this limitation and opens up a new perspective for the intelligent development of the flower and fruit tea industry. This article reviews the latest progress in the application advantages of CNN in the flower and fruit tea industry. A systematic review and meta-analysis were conducted on applying CNN in pest control, harvesting, and processing methods of flower and fruit tea raw materials (teas, flowers, fruits). Finally, an outlook was made on the relevant advanced progress and prospects. Compared with traditional machine learning methods, CNN has significant advantages in supervising flower and fruit tea production and processing. This review is expected to provide new insights into the application of intelligent technology in the tea industry.