Authors :
T.S.R. Krishna Prasad; P. Mahesh Kiran; M. Hima Sameera; P. Sri Sai Nachiketha; M. Mukesh Vamsi.
Volume/Issue :
Volume 8 - 2023, Issue 3 - March
Google Scholar :
https://bit.ly/3TmGbDi
Scribd :
https://bit.ly/3GfDvCc
DOI :
https://doi.org/10.5281/zenodo.7797284
Abstract :
- This research paper proposes a system for
detecting drones that use Raspberry Pi as its primary
computing platform and implements the SSD
MobileNetv2 architecture. The proposed approach
involves training a machine learning model using deep
learning and convolutional neural network algorithms.
The SSD Mobilenetv2 architecture is proposed due to
its accuracy and optimal performance in real-time
object detection. The dataset includes images of
numerous drones in various positions. The dataset has
undergone image augmentations such as flipping,
blurring, granulation and grayscale conversion, at
random, before training. Multiple cameras, connected
over a network, are connected to a Raspberry Pi
employing star network topology with Raspberry Pi as
the central hub. A dedicated machine, with the machine
learning model running on it, accesses the video feeds
from raspberry pi and infers them in real-time. The
detection results are sent to the raspberry pi. Computer
vision techniques are applied to the region of interest in
the video feeds to determine the drone's trajectory. The
system includes physical and digital alerts comprising
alarm systems and SMS alerts so that authorities can be
informed immediately whenever a drone is detected.
Keywords :
Drones, Raspberry Pi, SSD MobileNetv2, realtime detection, star network topology, trajectory,SMS alerts
- This research paper proposes a system for
detecting drones that use Raspberry Pi as its primary
computing platform and implements the SSD
MobileNetv2 architecture. The proposed approach
involves training a machine learning model using deep
learning and convolutional neural network algorithms.
The SSD Mobilenetv2 architecture is proposed due to
its accuracy and optimal performance in real-time
object detection. The dataset includes images of
numerous drones in various positions. The dataset has
undergone image augmentations such as flipping,
blurring, granulation and grayscale conversion, at
random, before training. Multiple cameras, connected
over a network, are connected to a Raspberry Pi
employing star network topology with Raspberry Pi as
the central hub. A dedicated machine, with the machine
learning model running on it, accesses the video feeds
from raspberry pi and infers them in real-time. The
detection results are sent to the raspberry pi. Computer
vision techniques are applied to the region of interest in
the video feeds to determine the drone's trajectory. The
system includes physical and digital alerts comprising
alarm systems and SMS alerts so that authorities can be
informed immediately whenever a drone is detected.
Keywords :
Drones, Raspberry Pi, SSD MobileNetv2, realtime detection, star network topology, trajectory,SMS alerts
