ShuffleDet: Real-Time Vehicle Detection Network in On-board Embedded UAV Imagery

[guanfuchen]

related paper

摘要

On-board real-time vehicle detection is of great significance for UAVs and other embedded mobile platforms. We propose a computationally inexpensive detection network for vehicle detection in UAV imagery which we call ShuffleDet. In order to enhance the speed-wise performance, we construct our method primarily using channel shuffling and grouped convolutions. We apply inception modules and deformable modules to consider the size and geometric shape of the vehicles. ShuffleDet is evaluated on CARPK and PUCPR+ datasets and compared against the state-of-the-art real-time object detection networks. ShuffleDet achieves 3.8 GFLOPs while it provides competitive performance on test sets of both datasets. We show that our algorithm achieves real-time performance by running at the speed of 14 frames per second on NVIDIA Jetson TX2 showing high potential for this method for real-time processing in UAVs.

[guanfuchen]

results

[guanfuchen]

architecture