Region and feature matching based vehicle tracking for accident detection
| Autor: | Vinod Karar, Sameer Suregaonkar, Shashi Poddar, Abhinav Saini, Neena Gupta |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
050210 logistics & transportation
Background subtraction Matching (statistics) Vehicle tracking system Computer science business.industry 05 social sciences Process (computing) Image processing 01 natural sciences Region of interest 0502 economics and business 0103 physical sciences Trajectory Computer vision Artificial intelligence business 010301 acoustics Feature detection (computer vision) |
| Zdroj: | IC3 |
| Popis: | Intelligent traffic monitoring using video surveillance is one of the most important aspects in administering a modern smart city. A recent growth towards machine learning and computer vision techniques has provided an added impetus towards this growth. In this paper, an image processing based vehicle tracking technique is developed that does not require background subtraction process to be applied for extracting the region of interest. Instead, a hybrid of feature detection and region matching approach is suggested in this article, which helps in estimating vehicle trajectory over consequent frames. Later, the tracked path is monitored for the occurrence of any specific event while the vehicle passes through an intersection. The proposed scheme is found to work promisingly on the real world dataset and is able to detect the occurrence of an accident between two vehicles. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|