Abstract
Drones have proven to be useful in many industry segments such as security and surveillance, where e.g. on-board real-time object tracking is a necessity for autonomous flying guards. Tracking and following suspicious objects is therefore required in real-time on limited hardware. With an object detector in the loop, low latency becomes extremely important. In this paper, we propose a solution to make object detection for UAVs both fast and super accurate. We propose a multi-dataset learning strategy yielding top eye-sky object detection accuracy. Our model generalizes well on unseen data and can cope with different flying heights, optically zoomed-in shots and different viewing angles. We apply optimization steps such that we achieve minimal latency on embedded on-board hardware by fusing layers, quantizing calculations to 16-bit floats and 8-bit integers, with negligible loss in accuracy. We validate on NVIDIA's Jetson TX2 and Jetson Xavier platforms where we achieve a speed-wise performance boost of more than 10x.
Abstract (translated)
无人驾驶飞机已被证明在安全和监视等许多行业领域都很有用,在这些领域,如机载实时目标跟踪是自主飞行警卫的必要条件。因此,需要在有限的硬件上实时跟踪和跟踪可疑对象。在循环中使用对象检测器,低延迟变得非常重要。本文提出了一种无人机目标检测快速、超精确的解决方案。我们提出了一种多数据集的学习策略,以获得最佳的天眼目标检测精度。我们的模型对看不见的数据进行了很好的概括,能够处理不同的飞行高度、光学变焦镜头和不同的视角。我们采用优化步骤,通过融合层、将计算量化为16位浮点数和8位整数,在嵌入式车载硬件上实现最小延迟,精确度损失可以忽略不计。我们在Nvidia的Jetson TX2和Jetson Xavier平台上进行了验证,在这些平台上,我们实现了超过10倍的速度性能提升。
URL
https://arxiv.org/abs/1904.02024