Efficient Object Detection Model for Real-time UAV Application

Unmanned Aerial Vehicles (UAVs) equipped with vision capabilities have become popular in recent years. Many applications especially been employed object detection techniques extracted from the information captured by an onboard camera. However, on UAVs requires high performance, which has a negative effect result. In this article, we propose deep feature pyramid architecture modified focal loss function, enables it to reduce class imbalance. Moreover, proposed method end model running UAV platform for real-time application. To evaluate architecture, combined our Resnet and MobileNet as backend network, compared RetinaNet HAL-RetinaNet. Our produced performance of 30.6 mAP inference time 14 fps. This result shows that outperformed 6.2 mAP.