Research On Real-time Face Recognition Technology Based On Drone Platform

Face recognition is one of the widely used authentication methods,especially in the field of contactless authentication with irreplaceable advantages.Obviously,when performing target tracking in an open outdoor location,the fixed face recognition mode has natural deficiencies such as monitoring blind spots and inability to maneuver,so the mobile mode must be used.The traditional mobile face recognition belongs to the ground maneuvering method,which can solve the above defects,but there are still problems such as building occlusion,limited plane maneuvering,and narrow field of view.Therefore,with the popularization and application of drones,the technology of high maneuver tracking and monitoring based on drones has become a research hotspot in recent years.This article focuses on the needs of multi-dimensional perspective tracking and identification of target people,rapid identification of targets in special environments and emergency situations,and compares how to build a flexible and portable intelligent drone platform and implement fast face recognition on mobile devices.discuss in depth.First of all,the commonly used ground-air master-slave face recognition work mode is studied and analyzed.Based on the concept of UAV intelligence,a fast face recognition based on UAV platform for ground-air collaboration and judgment is proposed.Operating mode.Secondly,in view of the limited computing resources and energy of drones,and the traditional face recognition methods with high complexity,weak feature learning initiative,and poor real-time performance,etc.,research and analysis of face detection and matching methods are compared and selected.Tiny YOLO implements face detection,selects FaceNet to achieve face matching,and in order to meet the requirements of its speed and duration on the drone platform,on the premise of ensuring the quality of face recognition,the above model is optimized.On this basis,in order to realize the real-time image processing of the UAV platform,the available hardware equipment is deeply studied,and the embedded vision equipment Raspberry Pi and Intel neural computing stick are selected to construct the UAV computing platform.Finally,the hardware and software design scheme of the UAV face recognition platform is given and implemented and tested.The test results show that the research results of this paper can work in a variety of working modes.Under the premise of ensuring the face recognition performance is unchanged or basically unchanged,the processing speed is increased or significantly increased to meet the requirements of real-time face recognition.