Year: 2025
Type: article
Abstract: In this paper, we propose a simplified architecture based on LSL-Net for detecting unmanned aerial vehicles, this architecture integrates the ability to track and detect them, based on convolutional neural networks. The biggest challenge is the creation of a model that allows us to have good results without the need for large computational resources, in addition to this we must take into account the scarcity of training images for drone detection. To address this problem, we developed simplified LSL-Net architecture using dilated convolutions that allows us to reach our objective. Finally, our proposed architecture integrates the detection and tracking of drones. Experiments show that our architecture performs well with limited resources, achieving an accuracy of 97\% in training and in the detection of new elements.
Source: Technologies
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $1,930
OpenAlex ID: W4410005275