Development of a Vision-Based Search System for Unmanned Aerial Vehicles

This paper presents the development of an autonomous unmanned aircraft system (UAS) that excelled in the Unmanned Platform Stereoscopic Cross-Domain Collaboration Challenge held in Yuxi, China, from August 1-15, 2022. The challenge was divided into four subtasks: (1)navigation and control, (2)recognition and localization, (3)online trajectory planning, and (4)task planning and scheduling, all of which were to be performed autonomously and safely by the robot. To meet these requirements, we designed and implemented a UAS with high autonomy and reliability, built on a hexacopter system, integrating sensors, an onboard computer, power batteries, and an electro-optical pod. We then developed the software algorithm and implemented the executable calculation code, integrated with the UAV hardware system. This paper outlines the main components and design process of the UAS and describes its application in a competition mission.