Method for pedestrian detection using ground plane constraint based on vision sensor
Changlong Jin, Xin Cui, Taekang Woo, Hakil Kim · 2016
Challenging problems pertaining to pedestrian detection still exist in Advanced Drive Assistance Systems. This paper presents a novel method to reduce erroneous pedestrian detection by moving vehicles. The algorithm utilizes camera parameters to calculate the distances between the moving vehicle and objects, and then builds ground plane constraints to detect real pedestrians. The proposed method not only reduces the number of false positives but also determines the distance of objects relative to the moving vehicle.