Multi-UAV Panoramic Video Surveillance System

As a new type of ground observation means,UAVs have been widely used in battlefield reconnaissance,environmental investigation,post-disaster investigation,and so on.Compared with traditional satellite system and other fixed shooting camera system,UAVs benefit a lot from its own motility and low-cost profile.However,single UAV’s capability is mainly limited by the flying altitude of the UAV,and the focal length of the camera,causing limited target area.This paper investigated the multi-UAV formation control,UAV-based panoramic video monitoring as well as the integrated framework to build up a panoramic video surveillance system,including the following key scientific problems:Firstly,the camera is firstly adjusted by Zhengyou Zhang calibration algorithm to obtain the suitable images for stitching.Then,the SURF algorithm is used to extract the feature points of the image,and the improved BFF(Best Bin First)algorithm is used to match the feature points.The method of checking the density degree and the RANSAC algorithm are used to filter the registration feature points,and the accuracy of the registration is effectively improved in the promise of small calculation burden.Then,the projection transformation matrix is calculated per the matching feature points,and the fusion of the gradient weights is used to smooth the fusion of images,which improves the visual effect of the fusion image.And based on the characteristics of UAV real-time video stitching,a dynamic key frame selection strategy is proposed.Finally,an image-based feedback control framework of multiple UAVs is proposed to achieve the most efficient video surveillance.In addition,the interface of the UAV video surveillance system is further developed,with the advantages of user-friendly design and simple operation.