To search, Click below search items. |
| All Published Papers Search Service |
|
|
|
Title |
Faraway Small Drone Detection based on Deep Learning |
|
Author |
Jae Kyun Kim, Young Jin Shin, Jin Ho Kim, Jung Kn Kim, Jae-Hoon Kim, and Seong Dae Lee |
| Citation |
Vol. 20 No. 1 pp. 149-154 |
|
Abstract |
The industry of small UAV (Unmanned Aerial Vehicle) such drones has recently been developing rapidly. The global UAV market is growing fast and is centered on the commercial and hobby drone markets, but it initially started with the military market. Due to the expansion of the drone market, related technologies including the miniaturization of drones and the increase in battery capacity have also been rapidly developing. As drone technology advances, the malicious use of drones including privacy violations and the transportation of small bombs have become serious problems. To prevent such problems, an anti-drone technology is also being developed. This paper carried out research on the anti-drone technology of drone detection. Studies on existing drone detection technologies have mainly been carried out through radars that converged with various sensors. The radar-based detection has problems such as the high cost and the need for an expert to operate it. Although studies using general web camera images without such radars have been conducted, they targeted only big and visually distinguishable drones. This study carried out exploration of the visually indistinguishable small drone images. The 2,085 images were used as a training set. 98.4% of the results were obtained through a mAP assessment of total 95-minutes video clip using the YOLOv3 model. |
|
Keywords |
Drone, Object Detection, Anti-drone, Deep Learning |
|
URL |
