In the subject of highway management, intelligent vehicle system and traffic management are beco... more In the subject of highway management, intelligent vehicle system and traffic management are becoming increasingly crucial in today’s time. As various vehicles are seen on the street daily, the number of these vehicles is increasing at a high rate. With this huge amount of vehicles, detection and tracking become a difficult job especially in developing and underdeveloped countries. With this thought, detecting and tracking vehicles from video frames is discussed and addressed in this work. In the realm of object detecting and tracking, the deep learning method has been widely applied. This research work proposes video-based vehicle detection and tracking. A new high definition highway vehicle data set containing around 8,000 images extracted from videos with proper annotation is made from the perspective of Bangladesh which provides a complete data foundation for vehicle detection and tracking based on deep learning. For making this data, different types of image processing methods had been used. For detection purposes, the YOLO v5 model, which is the most latest version of the YOLO model is being used; Also the MASK R-CNN model and SSD(Single Shot Detection) have been used for detection purposes. For tracking, we have used the YOLO v5 model, DeepSORT framework and GOTURN method. After detection and tracking, vehicle count and speed estimation are done. For vehicle counting and speed estimation, the YOLO v5 model is used. Classification of the vehicle is done in 10 categories. This information will aid in determining the priority and maximum users of a route and designing traffic patterns that will benefit the most people. Several highway surveillance videos based on different scenes are used to verify the proposed method
In the subject of highway management, intelligent vehicle system and traffic management are beco... more In the subject of highway management, intelligent vehicle system and traffic management are becoming increasingly crucial in today’s time. As various vehicles are seen on the street daily, the number of these vehicles is increasing at a high rate. With this huge amount of vehicles, detection and tracking become a difficult job especially in developing and underdeveloped countries. With this thought, detecting and tracking vehicles from video frames is discussed and addressed in this work. In the realm of object detecting and tracking, the deep learning method has been widely applied. This research work proposes video-based vehicle detection and tracking. A new high definition highway vehicle data set containing around 8,000 images extracted from videos with proper annotation is made from the perspective of Bangladesh which provides a complete data foundation for vehicle detection and tracking based on deep learning. For making this data, different types of image processing methods had been used. For detection purposes, the YOLO v5 model, which is the most latest version of the YOLO model is being used; Also the MASK R-CNN model and SSD(Single Shot Detection) have been used for detection purposes. For tracking, we have used the YOLO v5 model, DeepSORT framework and GOTURN method. After detection and tracking, vehicle count and speed estimation are done. For vehicle counting and speed estimation, the YOLO v5 model is used. Classification of the vehicle is done in 10 categories. This information will aid in determining the priority and maximum users of a route and designing traffic patterns that will benefit the most people. Several highway surveillance videos based on different scenes are used to verify the proposed method
Uploads
Drafts by Sabbir ahmed
on the street daily, the number of these vehicles is increasing at a high rate. With this
huge amount of vehicles, detection and tracking become a difficult job especially in
developing and underdeveloped countries. With this thought, detecting and tracking
vehicles from video frames is discussed and addressed in this work. In the realm of
object detecting and tracking, the deep learning method has been widely applied. This
research work proposes video-based vehicle detection and tracking. A new high definition highway vehicle data set containing around 8,000 images extracted from videos
with proper annotation is made from the perspective of Bangladesh which provides a
complete data foundation for vehicle detection and tracking based on deep learning. For
making this data, different types of image processing methods had been used. For detection purposes, the YOLO v5 model, which is the most latest version of the YOLO model
is being used; Also the MASK R-CNN model and SSD(Single Shot Detection) have been
used for detection purposes. For tracking, we have used the YOLO v5 model, DeepSORT framework and GOTURN method. After detection and tracking, vehicle count
and speed estimation are done. For vehicle counting and speed estimation, the YOLO
v5 model is used. Classification of the vehicle is done in 10 categories. This information
will aid in determining the priority and maximum users of a route and designing traffic
patterns that will benefit the most people. Several highway surveillance videos based
on different scenes are used to verify the proposed method
on the street daily, the number of these vehicles is increasing at a high rate. With this
huge amount of vehicles, detection and tracking become a difficult job especially in
developing and underdeveloped countries. With this thought, detecting and tracking
vehicles from video frames is discussed and addressed in this work. In the realm of
object detecting and tracking, the deep learning method has been widely applied. This
research work proposes video-based vehicle detection and tracking. A new high definition highway vehicle data set containing around 8,000 images extracted from videos
with proper annotation is made from the perspective of Bangladesh which provides a
complete data foundation for vehicle detection and tracking based on deep learning. For
making this data, different types of image processing methods had been used. For detection purposes, the YOLO v5 model, which is the most latest version of the YOLO model
is being used; Also the MASK R-CNN model and SSD(Single Shot Detection) have been
used for detection purposes. For tracking, we have used the YOLO v5 model, DeepSORT framework and GOTURN method. After detection and tracking, vehicle count
and speed estimation are done. For vehicle counting and speed estimation, the YOLO
v5 model is used. Classification of the vehicle is done in 10 categories. This information
will aid in determining the priority and maximum users of a route and designing traffic
patterns that will benefit the most people. Several highway surveillance videos based
on different scenes are used to verify the proposed method