Due to an increase in the population, the number of vehicles on the road has been increased. Automatic vehicle detection from these traffic scenes can help with better traffic management. This paper proposes methodology to develop intelligent automatic vehicle recognition in real time surveillance videos and also perform license plate recognition by replacing the old traditional method of monitoring traffic areas by human operators. Here detection and classification of vehicles on road has been done by implementing Residual Neural Network50 (ResNet50) and You Only Look Once version5 (YOLOv5) detection techniques. Datasets for seven classes of vehicles have been collected and annotations and labeling steps are done using ROBOFLOW software. The obtained results indicate that accuracy was found to be better for the model trained with ResNet50 when compared to YOLOv5. License plate recognition was also performed on car images using Optical Character Recognition(OCR). Index Terms-YOLO- You Only Look Once, ResNet- Residual Neural Network, OCR- Optical Character Recognition
Categorization of On Road Automobiles using Deep Learning Approach
2022-04-07
1050055 byte
Conference paper
Electronic Resource
English
DEEP LEARNING APPLICATIONS IN AUTOMOBILES
TIBKAT | 2021
|Engineering Index Backfile | 1913
|Measurements of road resistance of automobiles
Engineering Index Backfile | 1910
|Automobiles in road supervision cost data
Engineering Index Backfile | 1916
|