Researches On High-quality Tea Flushes Identification And Centering Method For Intelligent Plucking

In recent years, the tea industry has been rapidly developed in our country, and theresearch for the famous tea also emerges as the times require.Nowadays, the plucking ofhigh-quality tea still relies on manual work, which conveys the problem of labor shortages andhigh labor costs. The tea leaf shape and the integrity of the blade are destroyed easily duringmechanized harvesting.Therefore,the tea flushes plucking robot with the ability to pluck theflushes automatically and selectively is an urgent issure to be studied.This paper focuses on the research of spring tea, which is part of the research on teaflushes plucking robot.And this paper is also to study and explore the famous tea new shootsrecognition and actually determine the center point of the shoots.The main research contents:Using a camera to capture images of Long jing tea in a complex natural environment inRGB space, YIQ space HSI space and YCbCr space based on color feature recognition of theimages of new shoots. And analyzing the images segmentation and filtering;Finally,wecombine the median filtering with the threshold segmentation in order to identify the newshoots better.This paper uses morphological method to carry out new shoots center recognition on thebasis of the above research.Morphological algorithms such as corrosion,expansion, openingand closing operations are used to recognize the new shoots.According to the principle ofbinary morphology, using edge detection and image skeletonization to extract characteristics.In the edge detection, using first order differential operator Roberts operator, Sobel operatorand Prewitt operator and second-order operator LOG operator for identification of new tip. Inthe study of skeletonization, extracting the skeleton of the tea to recognize the new tiprespectively in RGB and YIQ space, which is to facilitate subsequent centroid location.This article uses the basic qualitative method and the weighted qualitative method toconfirm the center of the new tip, in the range of allowable error, this article confirm the centerof the new shoots accurately.This article carry out exploratory study of determination flushes center pointer. Usingmorphological features through the center of edge detection and comprehensive analysis of theskeleton to confirm flushes center,innovatively; Using the weighted method to confirmtheflushes center,successfully.Using machine vision and digital image processing technology, this paper realizes the goalof the identification and positioning of new shoots in natural environment, which providestechnical support for the realization of the machanization harvesting of famous tea.