%0 Journal Article
%T 基于可视化监测的自动采果作业系统设计与控制研究<br>Study on Design and Control of an Automated Fruit Harvesting System Based on Visual Monitoring
%A 顾怡红
%A 胡长德
%A 黄鸿猷
%J Dynamical Systems and Control
%P 67-76
%@ 2325-6761
%D 2025
%I Hans Publishing
%R 10.12677/dsc.2025.142008
%X 本研究旨在开发一种可视化监测的自动采果作业系统&#65292;通过整合Pixy CMUcam5视觉模块、SG-90微型伺服马达、MG996伺服马达和Futaba S148连续旋转型伺服马达等关键组件&#65292;并利用Arduino Mega控制板实现机构间运动的精确协调。通过坐标转换方法简化了果实定位问题&#65292;并结合柔性旗鱼式爪指设计&#65292;研发了三爪式夹爪作为西红柿采收的工具。通过实验分析&#65292;确定1.40 kgf的平均穿刺力作为最大夹取力的基准&#65292;夹爪上贴附的薄膜式力量传感器能够直接测量手指与果实间的力量&#65292;模拟测试显示最大夹持力为0.70 kgf&#65292;且夹取试验结果表明夹取动作对番茄造成的损伤极小。此外&#65292;本研究还测试了三种不同摩擦力材料的拉力和拉拔力&#65292;为后续夹爪材料的选择提供依据&#65292;为自动化采摘技术的发展提供重要的理论和实践基础。<br>This study aims to develop an automated fruit harvesting system with visual monitoring capabilities by integrating key components such as the Pixy CMUcam5 vision module, SG-90 micro servo motor, MG996 servo motor, and Futaba S148 continuous rotation servo motor, and utilizing an Arduino Mega control board to achieve precise coordination of movements between mechanisms. The problem of fruit positioning was simplified using coordinate transformation methods, and a three-fingered gripper with a flexible flag fish-style claw design was developed as a tool for tomato harvesting. Experimental analysis determined an average penetration force of 1.40 kgf as the benchmark for maximum clamping force. The thin-film force sensor attached to the gripper can directly measure the force between the fingers and the fruit, and simulation tests showed a maximum clamping force of 0.70 kgf, and the results of the clamping test showed that the clamping action caused very little damage to the tomato. Additionally, this study tested the tensile and pulling forces of three different friction materials to provide a basis for the selection of gripper materials in the future, offering important theoretical and practical foundations for the development of automated harvesting technology.
%K 机器视觉&#65292
%K 人工智能&#65292
%K 图像处理&#65292
%K 自动化技术<br>Machine Vision
%K Artificial Intelligence
%K Image Processing
%K Automation Technology
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=110894