【作者】 李志鹏；

【导师】 李东升；

【作者基本信息】 东北林业大学 ， 机械设计及理论， 2011， 博士

【摘要】 随着人们对以蓝莓为代表的天然抗氧化水果带给人们的健康益处的认识越来越深,需求也与日俱增,蓝莓在食品、药品、化工等方面具有巨大的市场潜力,种植蓝莓的经济效益十分显著,扩大种植面积是一种必然趋势。由于我国完全依靠人工采摘,而且蓝莓成熟期短,采摘耗资大、效率低,常因无法及时收获造成浪费,蓝莓的规模种植受到限制。发展机械采摘,成为我国蓝莓产业的迫切需求和必然趋势。为实现蓝莓采摘的机械化和自动化,需要研究一种高效智能的蓝莓采摘机械,并解决关键技术问题。本文重点解决两个问题,首先是采摘策略问题。要实现高效、安全地采摘,必须研究采摘策略,包括振动力大小和方向的确定、采摘机构运动的参数确定等,这不仅需要现场实测,据此建立蓝莓植株三维生长模型,还要在此基础上分析果枝的受力模型,建立蓝莓果树的振动模型,从而为实际采摘机的设计奠定基础。采摘机轨迹导航也是采摘成功的重要关键。采摘机由于单边牵引、地面凸凹和种植条件等原因,蓝莓采摘机车身沿种植垄行走时会发生偏斜,在垄距、株距固定的情况下,设计采摘机的空间受到限制,稍有偏斜就会造成蓝莓植株被压倒,甚至树根划伤。采摘机行走轨迹的规划引导是非常必要的。通过实测和抽象简化,对蓝莓植株的形态结构进行分析和建模,根据虚拟植物理论,以对树木进行动力学分析为目的建立植株模型。采用L系统理论,使用三维建模软件ProE建模,以VRML和OpenGL虚拟现实技术为手段,依据蓝莓树体结构特征,建立并实现蓝莓植株虚拟模型,实现蓝莓采摘过程可视化仿真。分析蓝莓采摘机振动系统呈现出的复杂、多自由度的振动性质,将蓝莓树枝简化为变截面连续振动系统,建立蓝莓植株的振动模型,并采用Matlab进行振动仿真,得到产生果枝分离的振动力来源。分析处于不同位置、不同角度枝桠上的蓝莓果实在采摘机工作时的受力情况,估算采摘成熟果实需求的外加采摘激振力,从而为采摘机采摘结构和工作参数的设计提供参考。根据蓝莓种植园的实地情况,确定识别和定位顺序,根据蓝莓特点研究蓝莓图像分割去噪、滤波、二值化等具体方法。利用图像分割的原理,采用Canny边缘检测得到蓝莓植株轮廓,得出的蓝莓植株中心线,调整蓝莓采摘机车体的进给朝向,进行轨迹引导和规划。根据成熟蓝莓颜色与背景颜色的明显差别评估当前蓝莓植株的熟果量,估算单株成熟蓝莓的多寡,藉此调整采摘机的采摘参数,确定自动采摘策略,实现智能采摘。根据蓝莓采摘机的结构特点和运动特点,分析行走偏斜和运动干涉产生的原因,根据种植条件和采摘机实际尺寸,得到采摘机轨迹干涉的边界条件和转弯安全距离的数学模型,实现采摘机的轨迹规划。更多还原

【Abstract】 As the health benefits form the natural antioxidant fruit such as blueberry are getting known, demand for the blueberries is also growing. Blueberry is widly used in food, medicine, chemical industry and other fields. So it is obvious that the economic benefits to cultivated blueberry is great, and it will be an inevitable trend to expand the planting area. Because the blueberry harvesting is depent on entirely manual labor in our country, and the harvesting season is short, picking cost is large and the picking efficiency is low, the economic losses are always causes because the fruits could not be harvested in time. The blueberry fields could not be expanse because the same reason. So developing picking machine is a urgent task in our country.In order to realize the mechanization and automation of blueberry picking, a kind of high efficient and intelligent blueberry picking machine should be studied, and its key technologies should be overcome. Two key technologies will be studied here, the first is the picking strategy. In order to realize efficient and safe picking, the picking strategy must be confirmed,which includes the vibration magnitude and direction,motion parameters of picking machine and so on. The3-d growth model of blueberry plants is establish with the measured data, and the mechanical model and vibration model of blueberry tree could also be established. All these will become a foundation for designing the picking machine. The second key technology is track navigation of the picking machine. Because of unilateral traction, the sunken ground, the specific planting condition and other reasons, a deflection occurs a lot when blueberry walking along the planting ridge. Because the distance of strain and the ridging are fixed, the picking space of the machine is restricted, a slight deflection may cause blueberry plants overwhelmed even roots scratch. It is very necessary to guide walking path of the picking machine.The morphological structure model of blueberry plant is analyzed and set up with the measured data and abstract simplification. According to the theory of virtual plant, the plant model is established on the purpose for trees dynamic analysis. With L system theory and3-d modeling software ProE, with the VRML and OpenGL virtual reality technology, and based on the characteristic of blueberry tree structure, the virtual model of blueberry plants is established and achieved, the visualization simulation of blueberry picking process is realized.After the vibration characterastics of the blueberry branches are analyzed, such as complex, multi-freedom, nonlinear, blueberry branches were simplified as variable cross-section continuous vibration system, the axial and radial vibration models are established, then the vibration models are simulated with Matlab, the source of vibration that separate the fruits and branchs is confirmed. The stresses on blueberry branches in picking machine work space are analysed in different position and different angle, the extra picking vibration need in picking ripe fruit is evaluated. Thus the design for picking machine structure and working parameters will be provided from these.According to the environment of the blueberry planting farm, the sequence of identifying and locating of blueberries and the trees is conformed, the methods for image segmentation, out-noising, filtering, binary are studied according to blueberry characteristics. The blueberry trees outline is got with the Canny edge detection according to the principle of image segmentation, then the centerline of the blueberry tree is got to heading the direction of the blueberry picking machine, so the track guide could be finished. According to the difference color between mature blueberry and background, the quantity of the ripe fruits of blueberry trees may be estimated, through which the picking parameters of blueberry machine could be adjusted, then the automatic picking strategies is confirm, and the intelligent picking propose is realized.According to the structure characteristic and movement characteristic of blueberry picking machine, the reasons of deflection and movement interference when the machine was walking were analyzed, and according to the planting condition and the size of the machine, the mathematical models for the track boundary conditions and the safe turning distance of the picking machine were got,and the picking track is planed and designed, a reasonable picking path of blueberry picking machine is achived.更多还原