基于单目视觉的智能物料分拣机器人的设计

为解决目前工业物料分拣机器人发展的需求问题,以模拟自动化物流系统的作业流程为目标,提出了一种基于单目视觉的智能物料分拣机器人的设计。为实现物料的自动分拣过程,以STM32作为核心控制器,驱动OV2640摄像头对图像进行实时采集、处理与分析,实现颜色识别和目标定位,并将处理结果传送给驱动控制系统。在图像标定方面,采用读取TFT屏中目标的坐标和求取目标的实际坐标,通过MATLAB软件来进行数据拟合,找出两个坐标的函数关系,从而实现目标的定位。采用遍历腐蚀算法、增量式PID算法、DBCSAN算法和Dijkstra算法,分别完成对随机摆放的物料的自动识别和定位、电机控制和路径规划。实验与竞赛结果表明,该单目视觉分拣机器人的分拣准确率和效率高,能够实现智能分拣功能。

Design of Intelligent Material Sorting Robot Based on Monocular Vision

In order to meet the demand of industrial material sorting robot,an intelligent material sorting robot based on monocular vision is proposed to simulate the operation flow of automatic logistics system.In order to realize the automatic sorting process of materials,STM32 is used as the core controller to drive the OV2640 camera to collect,process and analyze images in real time,realize color recognition and target positioning,and transmit the processing results to the driving control system.In image calibration,the target coordinates in the TFT screen are read and the actual coordinates of the target are obtained.Data fitting is carried out through MATLAB software to find out the functional relationship between the two coordinates,so as to realize the target positioning.Ergodic corrosion algorithm,incremental PID algorithm,DBCSAN algorithm and Dijkstra algorithm are used to respectively complete the automatic identification and positioning of randomly placed materials,motor control and path planning.The experiment and competition results show that the monocular vision sorting robot has high sorting accuracy and efficiency,and can realize intelligent sorting function.