Famous Tea Picking Mechanical Hand's Integrated Design Of Dsp Image And Control

The famous tea picking mainly depends on manual picking at this stage,the shortage of labor force,high labor costs,and low efficiency have become major factors restricting the production of famous tea.Therefore,using automatic picking robot combined with image processing and motion control is an inevitable trend for picking famous teas.At present,most of the image processing for tea is based on PC equipment,and PLC is used to control the robot movement,which make the equipment expensive and inconvenient to use.Using embedded technology to integrate tea image recognition and motion control,is a new way to solve the problems existing in tea picking,and it has great engineering significance for realizing tea picking automation.The main research contents of this paper are as follows:(1)According to the demand of picking robots,designing a hardware system based on embedded chips.The system consists of DSP,camera,servo driver and servo motor to integrate image recognition and motion control.(2)In the laboratory environment,tea image processing algorithms suitable for DSP are designed.Collect tea image information in laboratory lighting environment by using embedded vision device;Use the G component in tea color image as a grayscale image;Counting the gray values of fresh leaves and other backgrounds,choosing an appropriate threshold interval to binarize the grayscale image which can split fresh tea leaves;Using dilation and erosion algorithms to eliminate noise and interference;Using a improved binary domain image marking algorithm to calculate the number and location of fresh leaves,which can extract fresh leaf quantity information and location information.(3)Converting the image coordinate system to the actual robot coordinate system,then use the number information and position information of fresh leaves to control the movement of the motor according to the planned path by setting up DSP's ePWM module,and designed the entire picking motion control.(4)A lot of experiments were carried out in this paper,the statistics of the fresh leaf recognition speed and the recognition accuracy rate are made in image processing part.Analysis the unrecognized images to find out the reason,and the analysis of robot picking efficiency was done.This article uses embedded technology to integrate fresh leaf image processing and motor control on one embedded chip,achieving the recognition of fresh leaves and robot control.Feasibility experiment was done of famous tea automatic picking,and it can provide technical basis for the future research and development of intelligent picking equipment.