Centre de documentation
scientifique

Erfassung von Straßenverkehrsdaten mit elektro-optischem Sensor und automatischer Bildauswertung / T. Vogtle (1989)

Public Titre : Erfassung von Straßenverkehrsdaten mit elektro-optischem Sensor und automatischer Bildauswertung Titre original : [Enregistrement des Données sur le Trafic Routier par Capteur Electro-optique et Interprétation Automatique des Photos] Type de document : Thèse/HDR Auteurs : T. Vogtle, Auteur Editeur : Munich : Bayerische Akademie der Wissenschaften Année de publication : 1989 Collection : DGK - C Sous-collection : Dissertationen num. 352 Importance : 120 p. Format : 21 x 30 cm ISBN/ISSN/EAN : 978-3-7696-9400-0 Langues : Allemand (ger) Descripteur : [Vedettes matières IGN] Applications de télédétection [Termes IGN] acquisition de données [Termes IGN] caméra numérique [Termes IGN] capteur imageur [Termes IGN] détecteur à transfert de charge [Termes IGN] enregistrement de données [Termes IGN] étalonnage radiométrique [Termes IGN] photo-interprétation assistée par ordinateur [Termes IGN] système d'information routière [Termes IGN] temps réel [Termes IGN] trafic routier [Termes IGN] trafic urbain [Termes IGN] traitement parallèle Résumé : (Auteur) A still increasing number of vehicles and a higher demand for mobility at the same time require an optimization of traffic flow and most efficient use of existing road networks. For this a basic precondition is a fast, robust and sophisticated traffic data collection which allows traffic analysis, -forecast and -control. In this report such a system using an imaging sensor (electronic camera) and automatic digital image processing techniques is described. It has - if compared with classical methodes such as inductive loops or radar - some important advantages with regard to variety of collected data and flexibility in practical applications. For conception of a suitable hardware configuration fundamental components are presented and their qualification in the field of traffic data collection is discussed. A calibration of the system shows that the defined requirements can be fulfilled even with standard equipment. Because of rather different tasks the observations are divided in two parts, static and dynamic observations. The area of data collection is limited by the greatest attainable distance between camera and road on the one hand and by the geometric resolution on the other. The evaluation algorithms of traffic parameters utilize a priori information about the relevant objects and about general conditions in the field of traffic data collection. For detection of vehicles in static observations an extended approach of interframe differencing with a statistic base was created. The classification algorithm uses the determination of vehicle dimensions while position and velocity are measured by matching algorithms. The processing of dynamic scenes (aerial images) is realized by a consequent hierarchic approach. First - at the highest level - the segmentation of the road area will be carried out, then in a second step the 'areas of interest'- i.e. primitives like greyvalue-edges and -corners which could indicate vehicles are detected only in this part of the image. In a third step this few areas are examined to be a vehicle by a specific set of characteristics. With similar algorithms necessary control points can be located. For all measured values a detailed accuracy examination is carried out. In practical realisations a priori information can lead to a great amount of data reduction and therefore to a limitation of the necessary processing time. So for static observations a real time system was generated, while dynamic scenes need a higher level of computing power (for example parallel processing) than standard PC equipment can perform. During the first applications the system has proved itself with regard to accuracy and reliability. Because of its certain dependency on weather conditions the 24-hour operation has to be investigated, so this system can be used mainly as a very good supplement to classical methods. Numéro de notice : 58684 Affiliation des auteurs : non IGN Thématique : IMAGERIE Nature : Thèse étrangère Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=60416

Exemplaires(2)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
58684-01	35.46	Livre	Centre de documentation	Télédétection	Disponible
58684-02	35.46	Livre	Centre de documentation	Télédétection	Disponible