A Low-Cost and Fast Vehicle Detection Algorithm With a Monocular Camera for Adaptive Driving Beam Systems

The adaptive driving beam headlamp system is helpful to solve the traffic safety problems caused by abuse of high beams at night. fundamental problem for detect and track vehicles specific night environment light shape control. This paper proposes a new efficient algorithm detecting tracking based on video collected low-cost camera. related automobile regulations are analyzed classifying application scenarios summarizing performance requirements algorithm. A standard CMOS camera mounted behind windshield captures test from different scenarios. Some preprocessing approaches designed optimize captured so that can work independently color morphological characteristics rear lights utilized extract lamps vehicle. symmetry checked correlation coefficient method pair determine vehicle ahead preliminary. Then, Hungarian Kalman filter performed multiple occurrences in two consecutive frames correct detection results. Finally, an estimation given calculating position real world. experiments according referred obtained experimental results demonstrate get rates adaptability working condition system. Moreover, proposed has low time cost be applied embedded devices