基于OpenMV视觉系统的智能送药小车  

Smart Medicine Car Based on OpenMV Vision System

作　　者：付书添 查雪红 许超 Fu Shutian

机构地区：[1]江西理工大学,江西赣州341000

出　　处：《工业控制计算机》2022年第7期6-9,共4页Industrial Control Computer

摘　　要：自新冠肺炎疫情以来,“无接触”一词成了当下的热点词汇。针对在医院的“无接触送药”问题,设计了一种能自动送药的智能小车控制系统,该系统以STM32F407ZGT6作为主控制器,周围配备有OpenMV视觉系统、TCS34725颜色巡线系统和电机驱动系统,采用串级PID算法对特定颜色进行跟踪巡线,同时利用卷积神经网络对数字字模进行模型训练,OpenMV加载训练后的神经网络模型能够对数字字模进行数字识别。在起点给定目标数字后,小车根据目标病房号,识别地面的数字,到达终点后停车,等待取药后自动返回。经测试小车的稳定性高、鲁棒性好。Aiming at the problem of delivering the medicine without contact,this paper describes how to design an intelligent car system,which can automatically deliver the medicine.The system takes STM32F407ZGT6 as the main controller,and is equipped with OpenMV vision system,TCS34725 colorful line patrol system and motor drive system around.It adopts cascade PID control algorithm to track specific color line,meanwhile convolutional neural networks is used as model training to number.After loading the trained neural network,OpenMV can identify the number model.At the beginning,the target number is inputted into the car.It can identify the number on the ground,and stop in front of destination by number recognition.

关 键 词：STM32 机器视觉 串级PID 颜色巡线 

分 类 号：TP23[自动化与计算机技术—检测技术与自动化装置] TH788[自动化与计算机技术—控制科学与工程]