CMRT09

Stilla, Uwe

Bibliographic data

Monograph

Persistent identifier:: 856955019

Author:: Stilla, Uwe

Title:: CMRT09

Sub title:: object extraction for 3D city models, road databases, and traffic monitoring ; concepts, algorithms and evaluation ; Paris, France, September 3 - 4, 2009 ; [joint conference of ISPRS working groups III/4 and III/5]

Scope:: X, 234 Seiten

Year of publication:: 2009

Place of publication:: Lemmer

Publisher of the original:: GITC

Identifier (digital):: 856955019

Illustration:: Illustrationen, Diagramme, Karten

Language:: English

Usage licence:: Attribution 4.0 International (CC BY 4.0)

Publisher of the digital copy:: Technische Informationsbibliothek Hannover

Place of publication of the digital copy:: Hannover

Year of publication of the original:: 2016

Document type:: Monograph

Collection:: Earth sciences

Chapter

Title:: REFINING CORRECTNESS OF VEHICLE DETECTION AND TRACKING IN AERIAL IMAGE SEQUENCES BY MEANS OF VELOCITY AND TRAJECTORY EVALUATION D. Lenhart, S. Hinz

Document type:: Monograph

Structure type:: Chapter

In: Stilla U, Rottensteiner F, Paparoditis N (Eds) CMRT09. IAPRS, Vol. XXXVIII, Part 3/W4 — Paris, France, 3-4 September, 2009 183 3.3 Knowledge representation of traffic situations In order to evaluate the velocities of the vehicles we need to formulate our knowledge and expectations about the typical traffic situations such as “free flowing traffic”, “congestions” or “traffic jams”. In dependence of the state of traffic and the location with respect to intersections or traffic lights, different interactions between vehicles occur. A well substantiated statistical concept like Bayes’ theorem would provide a sound basis for evaluation. However, determining the probability density functions is hardly feasible because extensive and sufficient samples are missing. Hence, it is advisable to avoid a concept that claims statistical integrity. In contrast to Bayes’ theorem, fuzzy logic offers an intuitive method to represent knowledge of classes by easy parameterization (Zadeh, 1965). It is also frequently applied for modeling car following behavior (Brackstone and McDonald, 1999). Therefore, we decided to use fuzzy logic to describe our knowledge about traffic. 3.3.1 3D fuzzy membership function for active vehicles Let us define a fuzzy set A that describes vehicles which are actively involved in traffic. Besides normally moving cars, these may be standing vehicles in traffic jams or waiting at red traffic lights or other crossroads. Since we are only interested in the possibility of an object belonging to A, we neglect the alternative set A of inactive objects which may be false alarms of the detection, parking vehicles or erroneous tracks. For the fuzzy set A, a membership function needs to be defined, indicating the possibility ¡i A that a car belongs to A in dependence of its velocity v. However, ¡.i A also strongly depends on the traffic density D and the distance d from intersections. It is quite obvious, that in a free flowing situation in the middle of road segment the possibility that a car stands still is 0. In contrast to that, zero speed has a rather high possibility near intersections or in jam situations. In order to meet these different traffic situations, we have to consider the conditional possibilities fi A (v\D,d). In the sequel, the units for the measures given shall be v [km/h], D [cars/km per lane] and d [m]. First, we should outline the ranges of D and d where pi A may change significantly. A density of lower or equal to D = 30 corresponds to free flowing traffic while a density of D = 180 represents the maximum density of a traffic jam when there is almost no motion at all (Hall, 1999). Below 30, ii A (v,D\d) remains constant. The interesting range for d is approximately between 150 meters before an intersection because this describes the range where drivers start to brake and 50 meters behind the intersection where drivers accelerate until they reach their desired travel speed. Outside of the range of [-50m; 150m], /u A (v,d\D) is constant for all values of d. Over the entire space of v, D and d, this results in a 3D membership function fj A (v,d,D). Please note that the values also depend on the road type, speed limits intersection layout. For different road conditions, different functions have to be developed. The mentioned example function refers to a major city road with multiple lanes with a speed limit of 60 km/h and an intersection with a road of equivalent type controlled by traffic lights. D=30 and d=150 Figure 3: ID membership functions given a traffic density D = 0 and D = 180 respectively and distance d = 150 with support points (black dots) Distance d=5 Velocity v [km/h] Figure 4: 2D membership function at distance d=5, with ID support functions (black lines) To create this 3D function, support points have to be selected. I.e., given an open traffic situation (D = 30) and long distance from an intersection (d = 150) the possibility of a vehicle moving with a speed between 0-20 km/h in shall be 0, while the possibility at the same position in the same traffic situation shall be 1 for velocities between 50-70 and becoming 0 again at v = 100.

Access restriction

Copyright

fullscreen: CMRT09

Monograph

Chapter

Table of contents

Cite and reuse

Monograph

Chapter

Image

Image fragment

Citation links

Citation recommendation