【作者】 何金伊；

【导师】 杨福增；

【作者基本信息】 西北农林科技大学 ， 农业电气化与自动化， 2011， 硕士

【摘要】 近年来,机器人技术和视觉导航技术发展迅速,在农业领域中的拓展得到了广泛的关注。对农用视觉导航机器人技术进行研究与应用,不仅能够克服农业环境中的作业困难,降低劳动强度,提高经济效益,而且对加快农业的现代化进程有着现实意义。本研究在“863”项目(2008AA100903-7)资助下,以研制的新型履带式果园微耕机为对象,研究并设计了一套基于视觉导航的果园机器人控制系统。主要工作包括:(1)果园机器人控制系统的整体分析与设计。模块化地提出了果园机器人控制系统的体系结构并设计了其总体框架,将系统分为视觉子系统和运动控制子系统两个模块,对各子系统模块硬件进行选用与设计;(2)视觉子系统的设计。视觉子系统采用单目摄像头,经实验结果比较,选择基于OpenCV的改进两步法作为视觉子系统的标定方法,并深入地分析了果园机器人与导航路径的位置关系,推导出导航路径与预执行任务参数间的转换方程。采用高性能的TMS320DM642芯片作为处理的核心,根据两种规范化果园环境中的道路图像特点,详细地分析与设计了相适应的图像处理算法,并提出了基于两点法和预限制双边Hough变换法的中心线导航路径提取算法,分析了算法的稳定性和实时性;(3)运动控制子系统的设计。在运动控制子系统中,采用模糊控制法将导航路径参数(横向偏差和航向偏角)信息转换成相应的液压系统控制信号。使用MATLAB软件设计模糊控制器,并建立模糊控制系统模型,在SIMULINK环境中对模糊控制系统进行仿真,结果表明所设计的模糊控制器鲁棒强,准确性高,具有良好的控制效果,满足了设计需要;(4)果园机器人控制系统的软件实现。果园机器人控制系统的软件设计包括以DM642为处理核心的视觉子系统软件和以单片机AT89S52为控制核心的运动控制子系统软件两部分,对各部分中根据功能划分的子模块算法进行了详细的分析与设计,并给出了部分设计的程序。更多还原

【Abstract】 In recent years, the robot technology and visual navigation have developed rapidly, and its expansion in the agricultural field receives widespread concern. The research and application on the agricultural utility robot technology not only can solve the operational difficulties in the agricultural environment and reduce the labor intensity, but also suit the market demand, increase economic efficiency and has realistic meaning to accelerate the agricultural modernization. This paper is supported by the“863”project (2008AA100903-7) and its studying object is the new developed crawler orchard farming micro-machine. The paper researched and designed an orchard robot control system based on visual navigation. The main research work as followed:(1) Overall design and analysis of orchard robot control system. Firstly, proposing the orchard robot platform architecture and design the overall frame of control system hierarchically and modularly. Then, dividing the robot control system into vision subsystem and motion control subsystem, and designing the hardware and selecting components.(2) The design of vision subsystem. The vision subsystem has one single camera. The improved two-step calibration method based on OpenCV was chosen after experimental results compare of two methods. Based on analyzing in depth the position relationship between the orchard machine moving route and navigational path, the conversional equations of image information in the farming environment and planed mission parameters was derived. According the image features of the two standardized orchard road environment the article chosen, the images processing algorithm and the navigating center line extracted algorithm were analyzed and designed in detail with using the high-performance processing core of TMS320DM642.The extracting methods of navigation path based on two points and pre-limited bilateral Hough transform were proposed, and analyzed by stability and real-time.(3) The design of motion control subsystem. The motion control system uses the fuzzy control method to convert the lateral deciation and heading angle obtained by vision subsystem to the corresponded control signal of hydraulic system. The fuzzy controller and modeling of fuzzy control system were researched and designed detailed by using the software of MATLAB. Finally, fuzzy control system was simulated in the simulation environment of SIMULINK. The result shows that the fuzzy controller run with strong robust, response with high accuracy, has good control effect and meets the basic design request.(4) Software implementation of the orchard robot control system. The whole control system of the orchard robot was divided into the vision subsystem software which the control core was DM642 and the motion control system which the control core was AT89S52.Then, all the parts that were divided according to the functions of each subsystem were analyzed and designed detailed, and a part of procedures was given.更多还原