Drone vs. Bird Detection: Deep Learning Algorithms and Results from a Grand Challenge is a research report by Angelo Coluccia, Alessio Fascita, Arne Schumann, Lars Sommer, Anastasios Dimou, Dimitrios Zarpalas, Miguel Méndez, Daivd de la Iglesia, Iago, González, Jean-Phillipe Mercier, Guillaume Gagné, Arka Mitra, and Shobha Rajashekar.
The imperative to automatically detect and identify small drones has become crucial for various public and private stakeholders. This study introduces three distinct original approaches that participated in a grand challenge addressing the “Drone vs. Bird” detection problem.
The objective is to identify one or more drones within video sequences featuring potential distractors like birds and other objects, along with background or foreground motion. The algorithms are designed to raise alarms and provide accurate position estimates exclusively when drones are present, avoiding false alarms on birds or confusion with the surrounding scene.
The study focuses on three original approaches utilizing different deep learning strategies, evaluating their performance on a real-world dataset from the 2020 edition of the Drone vs. Bird Detection Challenge, provided by a consortium of universities and research centers. Results indicate varying difficulty levels among test sequences, influenced by factors like drone size and shape visibility, with sequences recorded by a moving camera and distant drones posing the most significant challenges. The performance analysis highlights complementary strengths among the different approaches, considering correct detection rate, false alarm rate, and average precision.
Publication Date– April 2021
Drone vs. Bird Detection: Deep Learning Algorithms and Results from a Grand Challenge contains the following main sections:
- Introduction
- Drone vs. Bird Challenge 2020
- Related Work
- Gradient Team
- EagleDrone Team
- Alexis Team
- Performance Comparison
- Conclusions
All articles published by MDPI are immediately available worldwide under an open-access license. No special permission is required to reuse all or part of the article published by MDPI, including figures and tables. Any part of the article may be reused without the consent for articles published under an open-access Creative Common CC BY license, provided the original article is clearly cited. For more information, please refer to https://www.mdpi.com/openaccess.
C-UAS Hub does not own this content and provides a link for users at the bottom of the page to access it in its original location. This allows the author(s) to track important article metrics related to their work. All credit goes to its rightful owner.
Authors-Angelo Coluccia, Alessio Fascita, Arne Schumann, Lars Sommer, Anastasios Dimou, Dimitrios Zarpalas, Miguel Méndez, Daivd de la Iglesia, Iago, González, Jean-Phillipe Mercier, Guillaume Gagné, Arka Mitra, and Shobha Rajashekar
See Also-
Passive bistatic radar: Target detection and interference
Drone Detection and Tracking Using RF Identification Signals
Deep Learning on Multi-Sensor Data for Counter UAV Applications
Post Image- Birds flying over Amman, Jordan (Image Credit: envatoelements by jancattaneo)