Over speed surveillance system using Deep Learning and Distributed
System
- Sanjog Gaihre,
- Aabhas Dhaubanja,
- Nabin Adhikari,
- Nitesh Das,
- Pratik Tamrakar,
- Binod Sharma
Sanjog Gaihre
Tribhuvan University Institute of Engineering
Author ProfileAabhas Dhaubanja
Tribhuvan University Institute of Engineering
Author ProfilePratik Tamrakar
Tribhuvan University Institute of Engineering
Author ProfileAbstract
This paper tackles worldwide road-safety and traffic-management issues
by implementing vehicle speed detection and license plate identification
technologies. A holistic method to improve road-safety and traffic
control addresses limited training data issues. The study stresses the
necessity for an efficient and reliable vehicle recognition, license
plate identification, and character segmentation system for precise
speed detection. A three-class Vehicle Detection model and customized
models for Numberplate detection, Character-segmentation, and
Character-Detection are presented to suit this need. Creating complete
training and testing datasets requires thorough data preparation, hand
clipping, and labeling. Data augmentation separates validation and
testing subsets while expanding the dataset. A robust and automatic
system for real-time vehicle speed detection and license plate
identification is the major contribution of this research. The suggested
system uses advanced deep learning to monitor and regulate traffic
efficiently, reducing manual intervention and improving road-safety.
Experimental findings reveal that the Vehicle Detection model can
recognize automobiles and the specialized models can detect license
plates, segment characters, and detect characters. The output of one
model feeds into the input of another on a distributed system, thus
these four models can operate simultaneously. These results demonstrate
the system's ability to improve road-safety and urban traffic
management.