一种视觉识别智能搬运机器人的设计

Design of intelligent transfer robot based on vision recognition

针对目前室内搬运机器人的移动精准度不高、视觉识别容错率低等问题,设计了一款以STM32F1和STM32H7为核心的双主控室内移动搬运机器人。采用13线霍尔编码器和增量式PID算法对电机进行闭环控制;采用电机里程计实现机器人定位,采用陀螺仪读取机器人姿态角度再利用PID算法对机器人空间定位和姿态进行矫正;利用OV7725摄像头获取图像,对图像进行中值滤波预处理,根据物体颜色并结合形状进行物料识别;利用PID算法对物料进行定位、抓取和放置。实验结果表明,在任意复杂的环境中,都能完成物料的正确抓取,并且能够精准地放置物料。

In order to solve the current indoor handling robots' problems,such low accuracy,and low visual recognition error tolerance,a dual-control indoor mobile transfer robot with STM32F1 and STM32H7 as the core is designed.It uses a 13-phase Hall encoder and incremental PID algorithm to control the motor in a closed loop,uses a motor odometer to achieve robot positioning,and uses a gyroscope to read the robot's posture angle and adopts PID algorithm to locate and correct the robot's space.Then,the OV7725 camera is used to get the image,perform median filter preprocessing on the image,recognize the material according to the color and shape of the object,and use the PID algorithm to locate the grab and place the material.The experiment results show that in any complex environment,can correct grasping of materials,and the materials can be accurately placed.