Object Classification with Roadside LiDAR Data Using a Probabilistic Neural Network

Object classification is important information for different transportation areas. This research developed a probabilistic neural network (PNN) classifier object using roadside Light Detection and Ranging (LiDAR). The objective was to classify the road user on urban into one of four classes: Pedestrian, bicycle, passenger car, truck. Five features calculated from point cloud generated LiDAR were selected represent difference between classes. A total 2736 records (2062 training, 674 testing) manually marked training testing PNN algorithm. data collected at three sites representing scenarios. performance evaluated by comparing result with those support vector machine (SVM) random forest (RF). comparison results showed that can provide highest accuracy among investigated methods. overall 97.6% database. errors in also diagnosed. Discussions about direction future studies provided end this paper.