基于集成设计的果蔬精准采摘实现方法研究

项目名称： 基于集成设计的果蔬精准采摘实现方法研究

项目编号： No.50805067

项目类型： 青年科学基金项目

立项/批准年度： 2009

项目学科： 轻工业、手工业

项目作者： 尹建军

作者单位： 江苏大学

项目金额： 20万元

中文摘要： 在非结构化环境下实现果蔬精准采摘是目前果实自动化收获领域的研究热点，也是农业装备智能化技术的发展趋势之一。本项目利用双目视觉伺服采摘机构并融合力反馈控制的方法，从机械系统集成设计角度探讨果蔬精准采摘实现的集成化方法，以解决视觉-机械系统关联的目标绝对定位问题和采摘中的机械损伤问题。从番茄的物理和机械特性着手，指导设计了柔顺夹持末端执行器的机械结构和传感控制系统。建立了两指无滑动稳定抓取力学模型，分析了采摘过程中引起番茄抓取损伤的影响因素，优化了柔性抓取的关键控制参数，提出了柔性抓取的辅助手段与节能控制措施。研究了基于切片透视的双目视觉测量模型，进行了眼在臂上的手眼标定，得到了考虑视觉定位与机械定位关联的目标绝对定位模型。研究了提取番茄与障碍物空间分布信息的图像融合方法和基于构形空间的关节型机械臂避障路径规划方法，避障采摘番茄的试验表明该方法可以应用于番茄的自动收获。开展了番茄采摘机械臂的无障碍快速运动路径规划和基于任务分解的多果顺序采摘试验。项目的研究成果可以为我国果蔬自动化采摘系统集成设计提供一定的理论参考和技术支撑。

中文关键词： 农业机械；集成设计；采摘机构；视觉定位；手眼协调

英文摘要： At present, it is a hot research topic in the field of automatic fruit-harvesting to realize precise fruit-picking under non-structuring environment, and is also one of development trends of intelligent technology of agricultural equipment. To solve absolute location problem of object and its mechanical damage problem, the project used the method that combined binocular vision servo picking mechanism with force feedback control of picking mechanism, and discussed the integration method of realizing precise fruit-picking from the point of integrated design of machine system. From tomatoes physical and mechanical properties to begin, mechanical structure and sensor control system of compliant gripper was guided to design. Mechanical model of nonslipping stable grab of two fingers was built up, and influence factors resulted from grab damage of tomato were analyzed during picking tomato. Some key control parameters of flexible grab were optimized and the supplementary means and energy saving measures were presented. A kind of binocular measurement model based on fault slice perspective was researched. Hand-eye calibration was carried out when binocular cameras were configured on the arm of manipulator, and absolute location models of object were obtained in consideration of the correlation between vision location and machine location. Image fusion method of extracting spatial information between tomatoes and obstacles and obstacle-avoidance path planning method of joint robot arm for picking tomatoes were studied respectively, and the test of avoiding erect obstacle to pick tomatoes showed that they may be applied into automatic harvesting of tomato. Path planning of free-obstacle quich motion of arm for picking tomato and sequential picking tests of mutiple fruits based on task breakdowm were also implemented respectively. The achievements in the Project may provide a theory of reference and technical support for integration design of automation fruit-picking system in our country.

英文关键词： Agricultural machine; Integration design; Picking mechanism; Vision location; Hand-eye coordination