Vehicle Classification in Video Based on Shape Analysis
Can Nguyen Van, Cuong Nguyen Ngoc · 2014
This paper aims at presenting some methods of representing image's features that help detect and classify vehicles from video. Proposed methods include: Method of representing shape, contour of vehicle or block of vehicle that can be classified. Paramaters of the Image's length in combination with parmaters of visual length of object that can used to classify object type or separate object. Use genaral deformable model of vehicle for allowing to be completely or partially occluding in the image. Apply some proposed methods of representing vehicle for vehicle recognition and classification system in traffic video. This paper also proposes a general working frame for the video - based traffic density detection and vehicle classification system in observation region. System was experimentally installed and obtained good results about the level of accuracy.