嵌入式采摘机器人视觉测量与避障控制系统研究

采摘机器人拥有自主收集信息并进行有效判断的能力,可以独立完成对果实的采摘作业,对满足水果种植需求、减小水果种植的劳动力投入及降低生产成本有着很重要的实际应用价值。为此,以嵌入式ARM智控系统为基础平台,设计了采摘机器人视觉测量与避障控制系统。该系统集机器视觉、视觉传感感知、伺服电机驱动和ARM智控模块于一体,建立了采摘机器人采摘运动学的数学模型,并通过BP网络神经型迭代学习算法测量果实的距离和球心坐标,对成熟果实进行精准识别和定位采摘。试验结果表明:采摘机器人能准确地进行自主采摘,成功率比较高,躲避障碍物的能力很强,更适合在复杂未知的果园中进行收获作业。

Research on Vision Measurement and Obstacle Avoidance Control System of Picking Robot Based on Embedded System

Harvesting robot has the ability to collect information and carry out effective judgment .It can be used to com-plete the picking of fruit .It is very important to have a very important practical application value for fruit growing de-mand, reducing the labor input and reducing production cost .This paper takes the embedded intelligence control system as the basic platform to design and research the picking robot vision measurement and obstacle avoidance control system , the system in machine vision , sensory perception , servo motor drive and arm intellectual control module in one , estab-lished picking picking robot kinematics mathematical model , and through BP neural network type iterative learning algo-rithm for measuring fruit distance and center coordinates , to mature fruits were accurate identification and localization of picking .The experimental results show that the picking robot can carry out self picking accurately , the success rate is relatively high , and the ability of avoiding obstacles is also very strong , and it is more suitable for harvesting operations in complex unknown orchard .