Research And Design Of Cutter Control Method For Tea Plucking Machine Based On Machine Vision

With the prosperity of the tea ceremony and the rapid development of tea drinks,the demand for tea is growing rapidly.Tea plucking has a stong seasonal characteristic,and its best time is very short.At present the work of tea plucking is mainly depend on manpower.However,the number of skilled workers is limited.It's a statistical fact that a tea garden with a thousand areas needs more than a thousand workers,which leads to prominent recruitment problems and expensive labor cost.Thus,the tea plucking machines,which can significantly improve efficiency,are needed urgently.The existing tea plucking machines are mainly adopted for the bulk tea,and cut the tea blindly,which lead to a low proportion of tea buds and can not meet the people's requirements for the tea quality.Consequently,it has great theoretical significance and application value to develop a cutter control method for tea plucking machine,which can cut tea selectively.To reduce the labor cost and improve the quality of tea cut for the tea plucking machine with crawler and ride,this thesis focuses on the research and design of the cutter control method based on machine vision.The main works and achievements of this thesis are as follows:(1)In order to keep the posture of cutter real-timely consistent with the top of tea plants and make the cutter be suitable for different topography and growth status of tea,the visual control method for crawler ride tea plucking machine are studied.First,by analyzing the fuselage structure and functional requirements of crawler ride tea plucking machine,the model of IPC and industrial camera are established.Then,the capture effects on different positions are compared,and the rear position is selected.At last,the algorithm library of HALCON is used to calibrate the camera internal parameters,which can eliminate the distortion of the image and improve the quality of the video.(2)For the sake of enhancing the color characteristics of tea and cutter positioning bar in the image,a set of tea image preprocessing methods are studied.First,the features of commonly used color space models are analyzed and compared.Then,contrast to result of intensity component method in the HSI color space,RGB independent component equalization method,three-dimensional joint equalization method in the RGB color space and multi-scale Retinex method,a color contrast enhancement method based on Lab color space is proposed.It uses the image dealed with RGB independent component histogram equalization as render chart to render a channel and b channel,and raises the contrast of L channel.This method can well enhance the color characteristics of tea buds and cutter positioning bar.(3)So as to accurately identify the area of tea buds in the image,a method of tea buds identifying based on RGB color space is proposed.First,the differences of R,G,B components are used to weaken the influence of light,which is usually uneven.Then,based on the color characteristics,OSTU is used to obtain the preliminary area of tea buds in the difference of G and B.Finally,for eliminating the mistaken areas and further identifying the final area of tea buds,a denoising method,based on the feature that mistaken areas are scattered and far away from the right,is proposed.(4)Aiming at solving the difficult problem of the positioning of cutter owing to its curve shape and thin cross-section,an indirect positioning method for curve cutter is designed.Meanwhile,by controlling the height of the left and right sides of cutter,the goal of keeping the posture of cutter real-timely consistent with the top of tea plants is achieved.First,for obtaining the relationship between reality and image quickly and accurately,a low-complexity and online camera external parameters calibration method is designed.Then,a positioning method with adaptive template for cutter positioning bar is proposed,which is based on its obvious color and shape.After that,the position of curve cutter in the image is calculated by the position of cutter positioning bar,the known structural parameters and the relationship between reality and image.At last,based on the relative position between the tea buds and the cutter,the height of left and right sides of the cutter are adjusted independently to keep the posture of cutter real-timely consistent with the top of tea plants.(5)By employing QT5 platform and the HALCON algorithms library,a software of cutter control for tea plucking machine based on machine vision is designed.First,the demand of software is analyed,including the functional demand and the performance demand.Then,the whole system is designed,including the architecture and the processing flow.Finally,the main function modules are realized,including the collection and display module,the automatic adjustment module and the interaction module.