基于STM32 + 百度云人脸识别的智能外卖背包系统设计

Design of Intelligent Takeout Backpack System Based on STM32 + Baidu Cloud Facial Recognition

摘要: 智能外卖背包终端系统采用多核处理器STM32MP157的Cortex-A7核运行Linux操作系统，并通过QT实现了人机交互界面，使用TCP向服务端上传人脸图片，并接收返回的人脸识别结果；Cortex-M4核执行例如温湿度检测、背包门开关控制等实时任务，并使用MQTT协议与华为云IOT服务器进行上、下行通信。Cortex-A7和Cortex-M4这两个核心基于OpenAMP库中的跨处理器通信的协议和框架RPMsg进行异核通信。服务端采用MVC架构、Spring Boot框架、Mybatis持久层框架构建，并通过百度人脸识别API提供人脸认证服务；Android客户端采用MVP架构构建，采用Retrofit + RxJava + OkHttp网络请求框架实现与服务端网络交互，Room框架实现客户端数据缓存。通过外卖背包设备终端，Android客户端，Spring Boot服务端，以及华为云IOT服务器搭建整个系统，实现用户对设备控制与监测，向用户提供较为完善的服务。

Abstract: The intelligent takeaway backpack terminal system uses the Cortex-A7 core of the STM32MP157 multi-core processor to run the Linux operating system, and implements a human-computer interaction interface through QT. It uploads facial images to the server using TCP and receives the returned facial recognition results; The Cortex-M4 core performs real-time tasks such as temperature and humidity detection, backpack door switch control, and uses MQTT protocol to communicate with Huawei Cloud IOT servers in both uplink and downlink directions. The Cortex-A7 and Cortex-M4 cores are based on the cross processor communication protocol and framework RPMSg in the OpenAMP library for cross core communication. The server is built using MVC architecture, Spring Boot framework, Mybatis persistence layer framework, and provides facial authentication services through Baidu facial recognition API; The Android client is built using the MVP architecture and utilizes the Retrofit + RXJava + OkHttp network request framework to interact with the server network. The Room framework is used to catch client data. By using the takeaway backpack device terminal, Android client, Spring Boot server, and Huawei Cloud IOT server to build the entire system, users can control and monitor their devices, providing them with comprehensive services.