Robust Real-time Pedestrian Detection in Aerial Imagery on Jetson TX2

TL;DR

This paper presents a lightweight, real-time pedestrian detection framework using YOLO on Jetson TX2, achieving high accuracy and speed for aerial drone imagery without additional training.

Contribution

It introduces a YOLO-based detection framework optimized for embedded systems that operates in real-time with high accuracy without requiring retraining.

Findings

Over 5 FPS detection speed on Jetson TX2

Achieves approximately 81 mAP on UAV pedestrian detection

Uses pre-trained models with minimal tuning

Abstract

Detection of pedestrians in aerial imagery captured by drones has many applications including intersection monitoring, patrolling, and surveillance, to name a few. However, the problem is involved due to continuouslychanging camera viewpoint and object appearance as well as the need for lightweight algorithms to run on on-board embedded systems. To address this issue, the paper proposes a framework for pedestrian detection in videos based on the YOLO object detection network [6] while having a high throughput of more than 5 FPS on the Jetson TX2 embedded board. The framework exploits deep learning for robust operation and uses a pre-trained model without the need for any additional training which makes it flexible to apply on different setups with minimum amount of tuning. The method achieves ~81 mAP when applied on a sample video from the Embedded Real-Time Inference (ERTI) Challenge where pedestrians are monitored by a UAV.