Development Of Vehicle Tracking And Counting System From Traffic Surveillance Video
A vehicle counting and tracking system automatically detects and classifies vehicles from traffic surveillance video sequences. The system are used to replace manual labor to collect vehicles data for various application such as transportation planning and road safety evaluation. The existing Vehicl...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2015
|
| Subjects: | |
| Online Access: | http://eprints.usm.my/40817/1/KUEH_CHIUNG_LIN_24_pages.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | A vehicle counting and tracking system automatically detects and classifies vehicles from traffic surveillance video sequences. The system are used to replace manual labor to collect vehicles data for various application such as transportation planning and road safety evaluation. The existing Vehicle Detection and Classification System does not have tracking and counting
module implemented. Tracking is required to enable automatic vehicle count. The objective of this project is to develop and implement tracking and counting feature into the existing vehicle detection and classification system, assess the vehicle detection and classification with tracking and counting feature system performances, and select the optimal parameters for tracking and
counting module. Visual Background Extractor (ViBE) is used to extract the vehicles (foreground) from the traffic surveillance video sequences. Simple tracking and counting algorithm is used to track and count the detected vehicle. Histogram of Oriented Gradient (HOG) is used to extract features from the detected vehicle. Multi-class Support Vector Machine (SVM) is used to classify the detected vehicle into four classes, which are motorcycle, car, lorry, and non-vehicle. The system is evaluated using two video sequences which are 670 seconds long with total of 20100 frames. The overall system performance achieves 78.19 % and 88.14% for
vehicle detection and classification, respectively. |
|---|