根据大疆Tello无人机的计算资源,结合飞控提供的软件接口,利用微软认知服务设计并实现了一种运行于边缘设备上的图像识别系统。文中主要介绍了系统的整体设计、硬件组成、应用程序的设计和实现。同时,以一种公开的人脸数据集为例,进行模型的训练和部署,边缘设备可以将获取的无人机实时图像通过HTTP协议与模型进行交互,获得其识别结果,降低了对本地硬件资源的需求,提升了处理效率,增强了系统灵活性。

An image recognition system running on the edge device is designed and implemented based on the computing resources of DJI Tello UAV,combined with the software interfaces provided by flight control and Microsoft cognition services.This paper primarily introduces the overall design,hardware composition,software design,and system implementation.Simultaneously,by taking a public face data set as an example,the model is trained and deployed.The edge device can exchange the real-time image of the UAV with the model through HTTP protocol to obtain its classification and detection the results.The system can reduce the requirements on local hardware resources,improve processing efficiency,and enhance the flexibility.