A Real-time Computer Vision System for Measuring Traffic Parameters
CVPR '97 Proceedings of the 1997 Conference on Computer Vision and Pattern Recognition (CVPR '97)
Near-optimal sensor placements in Gaussian processes
ICML '05 Proceedings of the 22nd international conference on Machine learning
Near-optimal sensor placements: maximizing information while minimizing communication cost
Proceedings of the 5th international conference on Information processing in sensor networks
Robust message-passing for statistical inference in sensor networks
Proceedings of the 6th international conference on Information processing in sensor networks
Model-driven data acquisition in sensor networks
VLDB '04 Proceedings of the Thirtieth international conference on Very large data bases - Volume 30
Distributed Online Simultaneous Fault Detection for Multiple Sensors
IPSN '08 Proceedings of the 7th international conference on Information processing in sensor networks
Image analysis and rule-based reasoning for a traffic monitoring system
IEEE Transactions on Intelligent Transportation Systems
Detection and classification of vehicles
IEEE Transactions on Intelligent Transportation Systems
A bayesian network approach to traffic flow forecasting
IEEE Transactions on Intelligent Transportation Systems
Automatic traffic surveillance system for vehicle tracking and classification
IEEE Transactions on Intelligent Transportation Systems
Image and video processing in wireless sensor networks
Multidimensional Systems and Signal Processing
Hi-index | 0.00 |
Traffic camera has played an important role in enabling intelligent and real-time traffic monitoring and control. In this paper, we focus on establishing a correlation model for the traffic cameras' vehicle counts and increase the spatial-resolution of a city's vehicle counting traffic camera system by means of correlation-based estimation. We have developed two methods for constructing traffic models, one using statistical machine learning based on Gaussian models and the other using analytical derivation from the origin-destination (OD) matrix. The Gaussian-based method outperforms existing correlation coefficient based methods. When training data are not available, our analytical method based on OD matrix can still perform well. When there is only a limited number of cameras, we develop heuristic algorithms to determine the most desirable locations to place the cameras so that the errors of traffic estimations at the locations without traffic cameras are minimized. We show some improvements in the performance of our proposed methods over an existing method in a variety of simulations.