Abstract | ||
---|---|---|
At present, the demand for transportation is continuously increasing, and the consequent traffic congestion problem has become more and more prominent. How to automatically and timely detect vehicles to analyze road traffic information is an important issue for intelligent traffic monitoring systems (ITS). In some existing methods for to detect vehicles, real-time performance and precision cannot be taken into account at the same time. Hereby, a method of automatic vehicle detection, which has the high performance on real-time and precision, is proposed in this paper. This method improves the YOLOv2 framework model in following aspects: introducing a new loss function, expanding the grid size, and optimizing the number and size of anchors in the model to automatically learn vehicle characteristics. Compared with YOLOv2, YOLOv3 and Faster RCNN, both the precision and the real-time performance of this method are improved competitively. |
Year | Venue | Field |
---|---|---|
2018 | PRCV | Grid size,Monitoring system,Computer science,Road traffic,Real-time computing,Vehicle detection,Artificial intelligence,Deep learning,Traffic congestion |
DocType | Citations | PageRank |
Conference | 0 | 0.34 |
References | Authors | |
9 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Wei Yang | 1 | 20 | 15.87 |
Ji Zhang | 2 | 0 | 0.34 |
Zhongbao Zhang | 3 | 404 | 27.60 |
Hongyuan Wang | 4 | 2 | 4.42 |