Pedestrian crossing detection based on evidential fusion of video-sensors

Abstract

This paper introduces an online pedestrian crossing detection system that uses pre-existing traffic-oriented video-sensors which, at regular intervals, provide coarse spatial measurements on areas along a crosswalk. Pedestrian crossing detection is based on the recognition of occupancy patterns induced by pedestrians when they move on the crosswalk. In order to improve the ability of non-dedicated sensors to detect pedestrians, we introduce an evidential-based data fusion process that exploits redundant information coming from one or two sensors: intra-sensor fusion uses spatiotemporal characteristics of the measurements, and inter-sensor fusion uses redundancy between the two sensors. As part of the EU funded TRACKSS project on co- http://www.sciencedirect.com/science/journal/0968090X patterns obtained and leads to high detection rates of pedestrian crossings with multi-purpose sensors in operational conditions, especially when a secondary sensor is available.

A Coarse-to-Fine Framework for Multiple Pedestrian Crossing Detection

“…Salmane et al [17] explored a video-based railway-road safety system for detecting hazard situations at level crossings. Additionally, Boudet and Midenet [7] performed spatial measurements on areas along a pedestrian crossing using traffic-oriented video sensors, supporting the occupancy patterns to exploit redundant information conveniently for pedestrian crossing detection; they used an evidence-based data fusion process to obtain redundant information by combining intra-sensor or inter-sensor data. Chyi-Ren et al [18] used environmental feature vectors to detect only one crossing in an image and determine the crossing area based on the environmental parameters.…”

“…For pedestrian crossing detection, there are different experimental datasets due to different application scenarios and purposes [4,5,7]. These datasets are mainly built for the single pedestrian crossing detection.…”

“…For example, Se [5,6] detects pedestrian crossings using lines grouping characteristics and projective geometry constraints. Boudet and Midenet [7] use videos to detect pedestrian crossing patterns through the fusion of sensor data. Liu et al [4,8] use videos to automatically detect pedestrian crossing and analyze its impairment status based on classifiers.…”

Automatic Pedestrian Crossing Detection and Impairment Analysis Based on Mobile Mapping System

“…With rather good robustness for pedestrian crossing detection in various scenarios, their approach works perfectly with plentiful precise camera parameters and data of pedestrian crossing width, but not with images taken at any time by DVRs or mobile phones with unknown interior and exterior orientation. In the research conducted by (Boudet, 2009), the position of a pedestrian crossing is located, with data from traffic surveillance videos converged, by constantly monitoring pedestrians' moving trails. Detection results via this approach are immune to occlusion or barriers but are subject to surveillance coverage and camera angles and unable to estimate pedestrian crossing scenarios.…”

Automatic detection of zebra crossings from mobile LiDAR data

“…One example is the automatic inventory of zebra crossing from mobile LiDAR data. Some successful attempts using image processing for the detection of zebra crossings have been done, although an accurate geo-positioning of the elements cannot be achieved [14][15][16]. Direct detection of zebra crossing from mobile LiDAR data with the corresponding geo-positioning information is not found in the bibliography.…”