UAV detection using neural networks
The availability of unmanned aerial vehicles (UAVs) has led to a significant increase in the number of offenses involving their use. This makes the development of UAV detection systems relevant. Solutions based on deep neural networks show the best results in detecting UAVs on video. This article pr...
Enregistré dans:
| Auteurs principaux: | , , , , , |
|---|---|
| Format: | Article |
| Langue: | anglais |
| Publié: |
Yaroslavl State University
2024-06-01
|
| Collection: | Моделирование и анализ информационных систем |
| Sujets: | |
| Accès en ligne: | https://www.mais-journal.ru/jour/article/view/1853 |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1839573140389232640 |
|---|---|
| author | Maria D. Averina Olga Levanova Darya V. Grushevskaya Kirill A. Kukharev Dmitriy M. Murin Maksim A. Kalinin |
| author_facet | Maria D. Averina Olga Levanova Darya V. Grushevskaya Kirill A. Kukharev Dmitriy M. Murin Maksim A. Kalinin |
| author_sort | Maria D. Averina |
| collection | DOAJ |
| description | The availability of unmanned aerial vehicles (UAVs) has led to a significant increase in the number of offenses involving their use. This makes the development of UAV detection systems relevant. Solutions based on deep neural networks show the best results in detecting UAVs on video. This article presents a study of various neural network detectors and focuses on identifying objects as small as possible, up to the size of $4\times4$ and even $3\times3$ pixels. The work investigates architectures SSD (VGG16) and YOLOv3 and it's modifications. Precision and recall metrics are calculated separately for different intervals of the object areas. The best result have been shown by YOLOv3 model with bbox parameters chosen as the result of object sizes clustering. Small ($3\times3$ px) drones have been successfully identified with 76% precision and a very small recall of 26%. For objects between 10 and 20 pixels in area, the recall is 64% with an accuracy of 75%. For objects with an area more than 20px the recall is about 90%, the precision is 89%, and the F1 score is 90%. These results show that it is possible to recognize even $4\times4$ pixel drones, which can be used in video surveillance systems. |
| format | Article |
| id | doaj-art-059fc5b615dd4ec1a8d16d888fcd1ff6 |
| institution | Matheson Library |
| issn | 1818-1015 2313-5417 |
| language | English |
| publishDate | 2024-06-01 |
| publisher | Yaroslavl State University |
| record_format | Article |
| series | Моделирование и анализ информационных систем |
| spelling | doaj-art-059fc5b615dd4ec1a8d16d888fcd1ff62025-08-04T14:06:43ZengYaroslavl State UniversityМоделирование и анализ информационных систем1818-10152313-54172024-06-0131218219310.18255/1818-1015-2024-2-182-1931410UAV detection using neural networksMaria D. Averina0Olga Levanova1Darya V. Grushevskaya2Kirill A. Kukharev3Dmitriy M. Murin4Maksim A. Kalinin5P.G. Demidov Yaroslavl State UniversityP.G. Demidov Yaroslavl State UniversityP.G. Demidov Yaroslavl State UniversityP.G. Demidov Yaroslavl State UniversityP.G. Demidov Yaroslavl State UniversityNational Research University Higher School of EconomicsThe availability of unmanned aerial vehicles (UAVs) has led to a significant increase in the number of offenses involving their use. This makes the development of UAV detection systems relevant. Solutions based on deep neural networks show the best results in detecting UAVs on video. This article presents a study of various neural network detectors and focuses on identifying objects as small as possible, up to the size of $4\times4$ and even $3\times3$ pixels. The work investigates architectures SSD (VGG16) and YOLOv3 and it's modifications. Precision and recall metrics are calculated separately for different intervals of the object areas. The best result have been shown by YOLOv3 model with bbox parameters chosen as the result of object sizes clustering. Small ($3\times3$ px) drones have been successfully identified with 76% precision and a very small recall of 26%. For objects between 10 and 20 pixels in area, the recall is 64% with an accuracy of 75%. For objects with an area more than 20px the recall is about 90%, the precision is 89%, and the F1 score is 90%. These results show that it is possible to recognize even $4\times4$ pixel drones, which can be used in video surveillance systems.https://www.mais-journal.ru/jour/article/view/1853uav detection |
| spellingShingle | Maria D. Averina Olga Levanova Darya V. Grushevskaya Kirill A. Kukharev Dmitriy M. Murin Maksim A. Kalinin UAV detection using neural networks Моделирование и анализ информационных систем uav detection |
| title | UAV detection using neural networks |
| title_full | UAV detection using neural networks |
| title_fullStr | UAV detection using neural networks |
| title_full_unstemmed | UAV detection using neural networks |
| title_short | UAV detection using neural networks |
| title_sort | uav detection using neural networks |
| topic | uav detection |
| url | https://www.mais-journal.ru/jour/article/view/1853 |
| work_keys_str_mv | AT mariadaverina uavdetectionusingneuralnetworks AT olgalevanova uavdetectionusingneuralnetworks AT daryavgrushevskaya uavdetectionusingneuralnetworks AT kirillakukharev uavdetectionusingneuralnetworks AT dmitriymmurin uavdetectionusingneuralnetworks AT maksimakalinin uavdetectionusingneuralnetworks |