茶叶茶梗的计算机在线识别技术及其应用

TECHNIQUE FOR ON-LINE TEA-STALK DISTINCTION AND ITS APPLICATION

分析了茶叶茶梗在颜色、形状和重量方面存在的差异，提出一种茶叶茶梗计算机在线识别新技术。该技术采用颜色、形状和重量三要素，光学、机械分选相结合，实现对茶叶茶梗多维模式识别。分类阈值可根据茶叶品种和分级要求在线调整。利用茶叶自除尘机理解决了茶尘对光路的阻塞问题。分类效果不受环境温、湿度的影响。该技术可用来替代入工进行梗、叶分拣，分拣效率可达到人工的２０～３０倍。本文介绍了茶叶茶梗计算机在线识别技术的实现方法和应用实例。

In this paper,a new technique is advanced for computer on-line distinction of tea-leaf and tea-stalk based on analysis of the differences between tea-leaf and tea-stalk in color,shape and weight.The multidimensional pattern recognition are realized according to the three factors by optical and mechanical methods.The thresholds values of recognition can be adjusted on-line according to the tea variety and the classify requirements.The problem of tea-dust clog to the optoelectronic sensors has been resolved by removing tea-dust with tea sliding down.The distinction effect would not be affected by air temperature or humidity.This technique can be used to replace manual operation for separating tea-stalk in tea processing and its efficiency is 20~30 times over manual operation.The method of realizing this technique is presented and some results in practice are shown.