Applying a Dynamic Recognition Scheme for Vehicle Recognition in Many Object Traffic Scenes
Włodzimierz Kasprzak, Heinrich Niemann · 1996
An adaptive object recognition scheme for image sequences of many object scenes is described. The scheme is applied for traffic object recognition under ego--motion. The recursive estimation of object states is performed by an extended Kalman Filter with modified error estimation, which is a neural network learning process. This new feature allows to separate the judgment needed for selection of best measurement among competitive image segments and the measurement judgment required by the recursive estimator. 1 Introduction Road traffic control [6] and driver support [7] is an attractive application field for image sequence analysis systems A reliable obstacle detection and classification in images of many--object scenes is still a challenging problem [8]. The complex nature of the subject makes it necessary to apply a dynamic model--based image analysis scheme, constraining the classes of recognized objects [3]. Usually such scheme employs a Kalman filter (KF) for recursive estimati...