Detecting Footpath Trespassers and Preventing Accidents
Keywords:
Automation, Obstacles, Classification, Adaboost, Support Vector Machine, Pedestrian, DetectionAbstract
Vehicles will eventually be fully automated, according to the automotive industry. While the automation of vehicles might lighten the burden on humans, we must also take care to ensure that safety is not compromised. The goal of this project is to identify any pedestrians in front of the host vehicle. The idea of project is to increase the accuracy in detection of pedestrian with minimal reaction time. This approach mainly focuses on developing optimized algorithm to detect pedestrian with accurate distance with host vehicle to generate AEB,ACC warnings. Distance of the pedestrian is estimated with width based distance correction. Further by introducing different classifier models for different distances better system accuracy can be achieved. A tracker module is used to keep track of the detections and make sure there are no wrong detections.