Vehicle Pose Estimation Based on Object Contour

Object recognition and pose estimation are the most fundamental computer vision problems. The pose estimation is highly essential in autonomous driving. In this article, we will look into a method for assessing the position of vehicles based on the external contour of an object. Once the features have been defined, vehicle pose estimation can be considered as a problem of measuring the similarity between the features of the obtained image and the reference. In this work, the external contours of CAD models of vehicles at different viewing angles are taken as a reference. The test results, presented to illustrate the proposed approach, confirm the relevance of this method.