Real-Time Drone Detection in Event Cameras via Per-Pixel Frequency Analysis

Michael Bezick; Majid Sahin

cs.CVMar 9, 2026

Real-Time Drone Detection in Event Cameras via Per-Pixel Frequency Analysis

Authors: Michael Bezick, Majid Sahin

Organizations: Johns Hopkins University APL

Abstract

Detecting fast-moving objects, such as unmanned aerial vehicle (UAV), from event camera data is challenging due to the sparse, asynchronous nature of the input. Traditional Discrete Fourier Transforms (DFT) are effective at identifying periodic signals, such as spinning rotors, but they assume uniformly sampled data, which event cameras do not provide. We propose a novel per-pixel temporal analysis framework using the Non-uniform Discrete Fourier Transform (NDFT), which we call Drone Detection via Harmonic Fingerprinting (DDHF). Our method uses purely analytical techniques that identify the frequency signature of drone rotors, as characterized by frequency combs in their power spectra, enabling a tunable and generalizable algorithm that achieves accurate real-time localization of UAV. We compare against a YOLO detector under equivalent conditions, demonstrating improvement in accuracy and latency across a difficult array of drone speeds, distances, and scenarios. DDHF achieves an average localization F1 score of 90.89% and average latency of 2.39ms per frame, while YOLO achieves an F1 score of 66.74% and requires 12.40ms per frame. Through utilization of purely analytic techniques, DDHF is quickly tuned on small data, easily interpretable, and achieves competitive accuracies and latencies to deep learning alternatives.

Figures & tables

Figure 1: (a) Far distance drone test at 300m using the Navitar 100mm F2.8 lens. (b) Direct sunlight test demonstrating the wide dynamic range of event based vision. (c) Multi-drone test - both identified correctly. (d)-(e) Handheld shaky-cam examples. Algorithm is able to accurately isolate the drone from the background, while keeping latency to real-time speeds despite large event throughput. (f)-(h) Angle of elevation experiment at $0^{\circ}$ , $10^{\circ}$ , and $90^{\circ}$ respectively. At $0^{\circ}$ , drone is barely visible (due to pose and possibly the focus of lens) and not identified, but at $10^{\circ}$ , drone is accurately identified, showing a limitation of DDHF. At $90^{\circ}$ angle of elevation, the drone is most easily identified. (i) Power spectrum of pixel where two overlapping rotor paths projected onto. (j) Typical power spectrum of a single-rotor pixel.

Figure 2: Visualization of one frame, comparing the spectra of a drone pixel versus a car wheel pixel. Both spectral flatness and harmonic comb scores from DDHF are able to distinguish between the spectrum of a spinning tire versus a spinning rotor.

Figure 3: Demonstration of the interpretable tuning capabilities of DDHF. For row (a) , the default hyperparameters obtained from the Bayesian optimization were used. For row (b) , $\tau_{sf}$ was tuned from $0.89$ to $0.5$ . For row (c) , $\tau_{\mathrm{comb}}$ was tuned from $1.5$ to $6$ . In both tuning cases, the bounding box becomes far more selective, filtering out the power spectra of the moving cars and effectively isolating the drone. Interestingly, only the inner parts of the drone rotor at this pose appear to be identified.

Condition	Detection Ours			Detection YOLO			Latency (ms) Ours			Latency (ms) YOLO			n
	P	R	F1	P	R	F1	Median	IQR	p95	Median	IQR	p95
Stationary - 10 m	$99.58$	$99.58$	99.58	99.79	98.98	99.38	2.30	$0.65$	$4.10$	11.91	1.17	15.24	$979$
Slow - 10 m	$97.18$	$96.57$	96.87	92.60	97.13	94.81	2.60	$1.06$	$5.06$	12.44	1.18	16.67	$350$
Fast - 10 m	$100.00$	$93.24$	$96.46$	97.26	98.61	97.93	3.01	$1.84$	$7.49$	12.07	0.74	13.45	$77$
Multi-Drone - 15 m	$93.63$	$93.53$	93.58	100.00	5.47	10.37	1.85	$0.46$	$3.48$	11.57	1.17	14.69	$448$
Direct Sunlight	$100.00$	$99.42$	99.71	100.00	96.53	98.24	2.29	$0.31$	$2.80$	13.22	1.26	16.78	$173$

Table 1: Head-to-head detection and latency at IoU=0.5. Our method is shown to almost always achieve far higher average detection scores and significantly lower latencies. As each frame at 30 fps corresponds to 33.33 milliseconds, latencies are comfortably below this budget, indicating our method is likely sufficient for real-time operation on the tested hardware. Latency is measured in milliseconds per frame, and inference is performed on an NVIDIA L40-8Q GPU.

n

= number of frames tested. We also report metric averages, weighted by number of frames per experiment.

Panel of Figure 4 — Figure 4: Performance versus angle of elevation. (a) Accuracy measured by F1-score; (b) inference latency.

Parameter	Symbol	BO Range	Final Value
Spectral flatness threshold	$\tau_{sf}$	$[0.3,0.98]$	0.89
Comb SNR threshold	$\tau_{\mathrm{comb}}$	$[1.5,6.0]$	1.5
# comb harmonics	$M$	$[4,16]$	4
Harmonic half-bandwidth	$\Delta k$	$[0,4]$	3
Min. matched harmonics	$h_{\min}$	$[2,8]$	4
Minimum passes / window	$n_{\min}$	$[2,10]$	6

Table 2: Bayesian optimization search space and results. All hyperparameters saw Uniform priors, and we use a GP with the expected improvement (EI) acquisition function.

Explore similar work

ODD-SEC: Onboard Drone Detection with a Spinning Event Camera
Mar 6, 2026Kuan Dai, Hongxin Zhang, Sheng Zhong +1Aerial Object Detection Event-Based Object Detection
Neuromorphic Automation and Intelligence Lab (NAIL) at School of Artificial Intelligence and Robotics, Hunan University, Changsha, China.
End-to-End Real-Time Drone-Based Person Detection Framework Using Deep Learning
Jul 12, 2026Payel Sarmah, Ayush Ranjan, Piyush Kaushik Bhattacharyya +2Object Detection Real-Time Object Detection
Centre for Drone Technology Indian Institute of Technology Guwahati Guwahati, India · Dept. of CSE KIIT University Bhubaneswar, India · Drone LAB National Institute of Electronics and Information Technology Bhubaneswar, India
AE-UAV: An Air-to-Air Event-Based UAV Tracking Benchmark and a Real-Time Frequency-Domain Tracker
Jul 16, 2026Zixin Jiang, Bing He, Chaoran Xiong +3Event-Based Vision Visual Object Tracking
Department of Electronic Engineering, Rocket Force University of Engineering, Xi’an 710025, China · Shanghai Key Laboratory of Navigation and Location Based Services, Shanghai Jiao Tong University, Shanghai 200240, China · State Key Laboratory of Submarine Geoscience, School of Automation and Intelligent Sensing, Shanghai Jiao Tong University, Shanghai 200240, China

Mar 6, 2026J/K move · Enter open · S save

ODD-SEC: Onboard Drone Detection with a Spinning Event Camera

Kuan Dai, Hongxin Zhang, Sheng Zhong, Yi Zhou

Neuromorphic Automation and Intelligence Lab (NAIL) at School of Artificial Intelligence and Robotics, Hunan University, Changsha, China.

Aerial Object Detection Event-Based Object Detection Real-Time Object Detection Event-Based Vision

Open paper AskPDF arXiv

The rapid proliferation of drones requires balancing innovation with regulation. To address security and privacy concerns, techniques for drone detection have attracted significant attention.Passive solutions, such as frame camera-based systems, offer versatility and energy efficiency under typical conditions but are fundamentally constrained by their operational principles in scenarios involving fast-moving targets or adverse illumination.Inspired by biological vision, event cameras asynchronously detect per-pixel brightness changes, offering high dynamic range and microsecond-level responsiveness that make them uniquely suited for drone detection in conditions beyond the reach of conventional frame-based cameras.However, the design of most existing event-based solutions assumes a static camera, greatly limiting their applicability to moving carriers--such as quadrupedal robots or unmanned ground vehicles--during field operations.In this paper, we introduce a real-time drone detection system designed for deployment on moving carriers. The system utilizes a spinning event-based camera, providing a 360° horizontal field of view and enabling bearing estimation of detected drones. A key contribution is a novel image-like event representation that operates without motion compensation, coupled with a lightweight neural network architecture for efficient spatiotemporal learning. Implemented on an onboard Jetson Orin NX, the system can operate in real time. Outdoor experimental results validate reliable detection with a mean angular error below 2° under challenging conditions, underscoring its suitability for real-world surveillance applications. We will open-source our complete pipeline to support future research.

Abstract

Explore similar work

ODD-SEC: Onboard Drone Detection with a Spinning Event Camera

End-to-End Real-Time Drone-Based Person Detection Framework Using Deep Learning

AE-UAV: An Air-to-Air Event-Based UAV Tracking Benchmark and a Real-Time Frequency-Domain Tracker