Vehicle Detection Under UAV Based on Optimal Dense YOLO Method

Abstract

In this paper, a deep neural network model based on small target detection under UAV platform is designed. Due to the One-stage detection model like YOLO having novel structure and great industrial application potential, this paper proposes a new model of detection based on YOLOv2 structure. Faced with missed detection problem of small target, a series of improved schemes are proposed, which are suitable for small vehicles' detection under aerial view angle, and can achieve real-time detection, including dense topology and optimal pooling strategy.