Design and testing of a machine-vision-based air-blow sorting platform for famous tea fresh leaves production

Abstract

Famous tea is a pillar of China’s tea industry and must be processed using fresh leaves of consistent tenderness and size. However, equipment with high accuracy and efficiency are unavailable for sorting fresh tea leaves (FTLs). Therefore, in this study, we designed a machine-vision platform for sorting FTLs. This platform subjects FTLs to a horizontal tossing motion and change trajectory by blowing air, thereby achieving the sorting of different types of FTLs. The overall sorting scheme consists of classification scheme, control scheme, and process scheme. We found that different types of FTLs have different flight trajectories when subjected to horizontal tossing. And have different optimal classification features and sorting parameters. When sorting high-grade FTLs (low-grade FTLs), the classification feature used is area and mean saturation (equivalent diameter, major-axis length and G variance), the designed platform achieved a recognition rate, purity, selection rate, and integrity rate of 100% (98.5%), 94.32% (87.47%), 91.67% (90.67%), and 100% (100%), respectively. Under optimal parameters, the production efficiency of the platform reached 25 kg/h. Overall, the results of this study indicate that the designed platform can meet the small-scale production requirements for famous tea.