A method called video analysis based on full information matching(FIM)is proposed to quantify the dynamic process of vehicles in avoiding collisions in traffic conflicts. The key point of the method is to obtain the exact position of the designated vehicle
and to calculate the transient information of the vehicle including its velocity
acceleration
course angles and front-wheel steering angle. Then the actions of the driver can be estimated
such as accelerating
decelerating
and swerving. All these measurements are taken without the drivers' awareness. Full information matching and sub-pixel estimation are used to eliminate the influence of vehicle contour changes on precision of vehicle positioning which are caused by the change of vehicle's shadow. Also the problem of camera's low pixels is solved by using these methods. The precision of vehicle positioning can be up to 0.1 pixel. An adjusted transmission transformation is used to eliminate the influence of shooting angle. A field experiment is carried out to obtain the driver's actions including shifting and swerving. It is proved that the driver's actions achieved by the analysis of high-altitude photography accord with the actual recording actions
and slight swervings that couldn't be observed or recorded artificially are caught as well.
关键词
Keywords
references
CHIN H C, QUEK S T. Measurement of traffic conflicts [J]. Safety Science, 1997, 26(3): 169-185.
COIFMAN B, BEYMER D, MCLAUCHLAN P, et al. A real-time computer vision system for vehicle tracking and traffic surveillance[J]. Transportation Research: Part C, 1998, 6(4): 271-278.
HAN Xiaowei, LI Yanping, LI Junsheng, et al. An approach of color object searching for vision system of soccer robot[C]∥ Proceedings of the IEEE International Conference on Robotics and Biomimetics. Piscataway, NJ, USA: IEEE, 2004: 535-539.
WAKABAYASHI H, RENGE K. Vehicle tracking system using digital VCR and ITS application to traffic conflict analysis for ITS-assist traffic safety [C]∥ Proceedings of the 13th Mini-EURO Conference on Handling Uncertainty in the Analysis of Traffic and Transportation Systems. Bari, Italy: Euro Working Group on Urban Traffic and Transportation, 2002: 356-362.
YANG Jianguo, YIN Xuquan, FANG Li, et al. Moving vehicle detection and tracking based on video of self-adaptive contour matching[J]. Journal of Xi'an Jiaotong University,2005,39(4):351-355.
GUO Rongxin, YANG Bangcheng, CAI Guangchen, et al. Comparative analysis of the subpixel displacement calculation methods in the digital speckle measurement[J]. Journal of Kunming University of Science and Technology, 2005, 30(2):55-59.
MENG Libo, MA Shaopeng, JIN Guanchang. On the performance of the subpixel displacement estimations used in digital speckle correlation method[J]. Journal of Experimental Mechanics, 2003, 18(3): 343-348.