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

Abstract

ABSTRACT:Pedestrian crossing, as an important part of transportation infrastructures, serves to secure pedestrians' lives and possessions and keep traffic flow in order. As a prominent feature in the street scene, detection of pedestrian crossing contributes to 3D road marking reconstruction and diminishing the adverse impact of outliers in 3D street scene reconstruction. Since pedestrian crossing is subject to wearing and tearing from heavy traffic flow, it is of great imperative to monitor its status quo. On this account, an approach of automatic pedestrian crossing detection using images from vehicle-based Mobile Mapping System is put forward and its defilement and impairment are analyzed in this paper. Firstly, pedestrian crossing classifier is trained with low recall rate. Then initial detections are refined by utilizing projection filtering, contour information analysis, and monocular vision. Finally, a pedestrian crossing detection and analysis system with high recall rate, precision and robustness will be achieved. This system works for pedestrian crossing detection under different situations and light conditions. It can recognize defiled and impaired crossings automatically in the meanwhile, which facilitates monitoring and maintenance of traffic facilities, so as to reduce potential traffic safety problems and secure lives and property.

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

“…The proposed method tend to focus on whole regions of interest, but mainly focus on the vanishing points detection at the right directions in the image. When there is only one true pedestrian crossing in an image, we will summary much similar results with [4,6]. Experimental results reveal that the methods tend to fixate prior on clear objects in an image.…”

“…As shown in Table 3, the proposed method has a higher recall rate 96.85%, while preserves a higher precision rate 95.4%. Under the varying condition of illumination or orientation changes in camera images, unstable visual information and geometric transform can lead to imperfect experiment result for [4,5,12]. Meanwhile, the grouping geometric features in figure-ground segmentation tend to be fail when pedestrian crossing is occluded partially [6,14].…”

“…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.…”

“…With the advent of smart phones and location-based services, there has been an increased interest in providing route guidance to pedestrians under different environments [1][2][3]. As an important part of transportation infrastructures, pedestrian crossing serves to secure pedestrians' lives and possessions and keep traffic flow in order [4]. What's more, in the generation of walking paths, one of the major safety considerations is to ensure that streets are crossed at places with established pedestrian crossings.…”

“…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. A thorough literature review is provided in Section 2.…”