Vehicles detection in complex urban traffic scenes using a nonparametric approach with confidence measurement

Aiming to efficiently resolve the problem that the subtraction background model is easily contaminated by slow-moving or temporarily stopped vehicles in complex urban traffic scenes, a novel Nonparametric Approach with Confidence Measurement (NPCM) is proposed for vehicle detection in complex urban traffic scenes. According to the current traffic state, each pixel of the background model is assigned a confidence measurement. The foreground decision depends on an adaptive threshold and the background model update is based on whether the current pixel point is in confidence period. Using the real-world urban traffic videos, the overall results of the detection accuracy analyses demonstrate that the NPCM achieves better performance of quantitative evaluation than other state of the art methods. Not only the NPCM can accurately detect the slow-moving or temporarily stopped vehicles, but also the similarity and F-measures of the NPCM are over 0.839 and 0.912, higher than the other compared methods in traffic-light sequence of daytime, respectively. The experimental results show that the NPCM is effective and is suitable for the real-time implementation in vehicles detection of complex urban traffic scenes.