【作者】 张波；

【导师】 翟长远；

【作者基本信息】 西北农林科技大学 ， 农业电气化与自动化， 2017， 硕士

【摘要】 农药是农业生产过程中防治病虫草害的有效手段,在传统的大田喷杆式喷雾过程中,由于缺少实时追踪作物行的手段,而采用大流量、宽角度的喷头对多作物行进行全覆盖喷雾,使大量农药沉积到行间土壤中,造成严重的环境污染;同时,由于农村青年劳动力转移造成农村劳动人口缺失以及人工施药过程中农药对人体造成严重的危害,使发展无人喷雾作业系统显得尤为重要。本文基于GPS和机器视觉技术设计了田间自走式对行喷雾机器人,主要研究内容与结果如下:(1)搭建了四轮独立转向机器人移动平台,具有高地隙特点,采用四对直流电机进行移动平台运动和转向控制,每个电机由一个单独的电机驱动器进行控制,利用增量式编码器采集电机转速和偏转角度,实现移动平台运动速度和转向角度的准确控制,解决了二轮转向移动平台在进行地头转向时需要较大转向半径的问题。(2)采用DGPS和电子罗盘构建机器人田间自主导航控制系统,以机器人移动横向偏差和航向偏差作为输入量,以机器人转向电机角度作为控制量,基于模糊控制算法实现机器人田间直线导航,试验表明,机器人直线导航最大误差11.32cm。(3)基于机器视觉技术设计了喷杆式施药机对行喷雾控制系统,通过相机获取作物行RGB图像,设计了以垂直投影法为基础的作物行中心线提取算法,提高了处理效率和识别准确性,试验表明,作物行中心线提取算法平均耗时12.51ms,喷杆横向偏移量计算误差小于0.44cm。(4)基于AVR系列单片机ATMega16设计了电动喷杆控制系统,通过RS232串口接收上位机发送的喷杆横向偏移量,设计了延时策略和基于位移传感器的直流推杆电机位置控制方法,实现对电动喷杆位置的精确控制,试验表明,电动喷杆右移最大误差0.3cm,左移最大误差0.5cm。(5)以四轮驱动农用小车为平台开展了对行喷雾控制实验室和田间试验,试验表明,小车平均速度为0.26m/s时,对倾角为5°、10°和15°的模拟作物行最大对行误差分别为3.22cm、2.86cm和2.51cm;小车平均速度为0.24m/s时,车身最大偏移距离14.05cm,对田间玉米幼苗的最大对行误差为4.86cm。更多还原

【Abstract】 Pesticide is a kind of effective method to prevent plant diseases and pests.During traditional boom spraying in field,high-flow and wide-angle nozzles were wildly used to cover multiple crop rows because of lack of the method to follow the crop row in time.It caused a heavily environment pollution that much pesticide deposited to the soil between the crop rows.Meanwhile,more and more young labors in countryside moving to the city caused the lack of labors in countryside nowadays.And pesticide brought a great harm to human body when spraying artificially.So it was so significant to develop spray robot in filed.A autonomous navigation and row-follow spray robot was designed based on GPS and robot vision.The main research contents and results are as follows:(1)Four-wheel independent drive moving bench for robot was designed.It had the characteristics of high ground clearance.Four pairs of DC motor which was drove by motor driver were used for moving control and steering control.Incremental encoders were used to detect the speed and position of motors.Thus,a moving bench with precise moving speed and steering angle was finished.It solved the problem that tow-wheel steering bench was difficult to turn around at the end of field.(2)Robot navigation control system in filed was built based on DGPS and compass.Fuzzy control algorithm was designed to implement the straight line navigation of robot in field with robot lateral deviation error and the heading error as input,steering angle as the output.The tests indicated that the maximum error of robot straight line navigation is 11.32 cm.(3)Row-follow spray control system for boom sprayer was designed based on robot vision.Crop rows RGB images acquired by a CCD camera was analyzed by center line detecting algorithm which was optimized based on vertical projection.Processing time and accuracy rate were enhanced.The results showed that the modified algorithm of crop row centerline extracting cost 12.51 ms in average with calculation errors of 0.44 cm.(4)An ATMega16 controller for electric boom was designed to receive the parameters through an RS232 serial port form upper computer,and adjusted the electric boom to the left or right with movement monitoring by a displacement sensor.Experiments were conducted in the lab and field.The results showed that the modified algorithm of crop row centerline extracting cost 12.51 ms in average with calculation errors of 0.44 cm.The electric boom adjustment error was less than 0.3 cm when shifting right,and less than 0.5 when shifting left.(5)Some experiments were designed to detect the performance of the row-follow control system.The row-follow control error was less than 3.22 cm,2.86 cm and 2.51 cm under the speed of 0.26m/s,when following simulated crop rows with obliquity of 5°,10° and 15°,while less than 4.86 cm when following the rows of corn seedings with the speed of 0.24m/s and the shifting of 14.02 cm.It provided an effective solution for crop rows tracting and row-follow spraying.更多还原