3D objects modelling and description from a traffic flow

Contemporary research activities in the area of transportation systems usually utilize computer vision techniques. These activities are mostly oriented to the analysis of traffic flow. Key parts of the analysis consist of detection, extraction, modelling and recognition of objects in the traffic flow. Additional information can be obtained by the object tracking, too. The main aim of this contribution is to propose a new method for modelling 3D objects that are moving in the traffic flow. At the beginning of this paper, basic methods of object detection and extraction are shortly described. Moreover, the modified algorithm based on the object extraction and 3D models creation by mesh is also proposed. That mesh model is generated from a depth map. Finally, the results of proposed method are presented in the last part of this paper.