A vehicle occupant counting system using near-infrared (NIR) image

This paper investigates a method to detect and count occupants in vehicles. Proposed near-infrared (NIR) image capture method can effectively deal with the challenge due to poor light conditions, windshield reflection, tinted windows and shadows on windshields to improve the clarity of the captured image of the vehicle interior. Furthermore, we proposed an algorithm to detect the occupants in vehicles. Firstly, the vehicle windshield candidate regions are extracted based on mean-shift segmentation and Hough transform methods, we select the most adaptive candidate region from the candidate regions using proposed maximum energy method. Then, the occupants in the windshield region were segmented through Histograms of Oriented Gradients (HOG) descriptors based human detection method. The experimental results demonstrate the efficiency of our algorithm.