Urban target detection algorithm based on multi-modal fusion of UAV
Abstract
Urban target detection algorithm based on multi-modal fusion of UAV
Incoming article date: 27.07.2024The use of small drones to detect urban targets such as vehicles at low altitude in cities has gradually become a mainstream means. In view of the existing problems of low detection accuracy of single-mode detection network caused by visible light detection, inability to work at night and blurred edge of infrared detection targets in actual scenes, In this paper, a multi-modal UAV detection algorithm based on image fusion and deep learning network is proposed. Firstly, based on DUT-VTUAV visible-infrared registration data set and TIF image fusion algorithm, multi-mode fusion data set is constructed. Secondly, by comparing the detection accuracy, speed and number of parameters of the existing YOLO series network, the lightweight network YOLO v5n which is most suitable for the mobile deployment of UAVS is selected. Finally, a multi-modal fusion detection algorithm is formed by combining image fusion algorithm and target detection model. Comparative experiments on vehicle data sets show that compared with single-mode detection, the detection accuracy of the proposed algorithm is effectively improved, with mAP up to 99.6%, and a set of visible-infrared image fusion detection can be completed within 0.3s, which has high real-time performance.
Keywords: target detection, YOLO, multimodal fusion, data fusion, TIF algorithm