Intelligent Tea Sorting System Based on Computer Vision

In the inspect process of fresh picking tea, an essential segment is variety selection; And in the cultivation and research stage, variety research and identification is also a means to improve variety protection. Traditional methods often have problems such as a single way, not accurate segmentation and so on. With the rapid development of artificial intelligence and machine vision, the intelligent training models of deep learning are also becoming more and more widely used in the recognition and classification of objects. This thesis proposes a new tea sorting system based on CNN, which can extract characteristic information from shape and color, and can classify 10 kinds of tea, realizing the intelligent, rapid and accurate sorting of tea variety.