基于机器视觉的中耕除草机的研制与试验  

Development and experiment of machine-vision-based cultivator

作　　者：葛艳艳[1] 安秋[1] 张庆怡[1] 霍茂森 

机构地区：[1]南京农业大学工学院,江苏南京210031

出　　处：《甘肃农业大学学报》2015年第5期172-176,共5页Journal of Gansu Agricultural University

基　　金：中央高校基本科研业务费资助项目(KYZ201006);南京农业大学青年科技创新基金资助项目(KJ09030);江苏省大学生创新项目(JSS1239)

摘　　要：传统的中耕机械除草作业极易损伤作物,使其作业效果大打折扣.为了充分发挥中耕除草作业的优势,研制了一种基于机器视觉技术的新型中耕除草机.该机具采用双横梁摆动机构,并加装电液控制模块和作物行视觉检测模块.采用2层式控制结构,上位机实现作物行的视觉识别和定位算法,基于单片机的下位机控制器实现锄具的侧向运动控制,上、下位机间采用CAN总线通讯.通过测试表明,整个系统工作正常,锄具的侧向运动速度在50~75mm/s,作业精度小于15cm,可实现在无人参与下的行间除草作业要求.Traditional mechanical weeding for inter-tillage is not used wildly because of the damage to growing crops.In order to make better use of this method,a new inter-row cultivator was developed based on machine vision technology.Double beams swing mechanism was introduced,while adding an electro-hydraulic control module and a crop rows vision detection module.A two-layer control structure was adopted.Algorithm of the vision recognition and positioning were realized in PC.The control of hoe side-shift was done by the microcontroller-based salve controller.CAN bus was used to communicate between PC and slave controller.The experiments were carried out and the results showed that,the system worked properly,the inter-row precise weeding could be done in unsupervised manner,side-shift speed of hoes was between 50mm/s and 75mm/s,weeding accuracy was below 15 cm.

关 键 词：机器视觉 机械除草 作物行识别 中耕除草机 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置] S224.1[自动化与计算机技术—控制科学与工程]