Vision-Based Potential Pedestrian Risk Analysis on Unsignalized Crosswalk Using Data Mining Techniques

Abstract

Though the technological advancement of smart city infrastructure has significantly improved urban pedestrians’ health and safety, there remains a large number of road traffic accident victims, making it a pressing current transportation concern. In particular, unsignalized crosswalks present a major threat to pedestrians, but we lack dense behavioral data to understand the risks they face. In this study, we propose a new model for potential pedestrian risky event (PPRE) analysis, using video footage gathered by road security cameras already installed at such crossings. Our system automatically detects vehicles and pedestrians, calculates trajectories, and extracts frame-level behavioral features. We use k-means clustering and decision tree algorithms to classify these events into six clusters, then visualize and interpret these clusters to show how they may or may not contribute to pedestrian risk at these crosswalks. We confirmed the feasibility of the model by applying it to video footage from unsignalized crosswalks in Osan city, South Korea.

Analysis of Vehicle–Pedestrian Interactive Behaviors near Unsignalized Crosswalk

“…The proposed analytical system in this paper had three main objectives: (i) to automatically extract the trafficrelated objects' behavioral features which affect the likelihood of potentially dangerous situations after detecting them into individual objects; (ii) to analyze their behavioral features and relationships among them by camera location; and (iii) to support administrators making efficient decisions to improve the safety of the road environment. Based on the authors' previous studies (21,22), the scale was expanded to larger urban areas with more CCTVs available, as well as longer daily study periods, and various behavioral features which affect the potential pedestrian risks were analyzed. In addition, unstructured video footage was converted into structured-typed datasets with the design of the database schema to analyze potential traffic risks.…”

“…To determine ground tips for them, the object mask and central axis line of vehicle lane were identified, and then perspective transform was conducted using the ''transformation matrix'' function in OpenCV library. The more detailed procedures for these are explained in the authors' previous studies (21,22). This paper focuses on the object tracking and indexing process.…”

Street life and pedestrian activities in smart cities: opportunities and challenges for computational urban science

“…Travel trajectories have been used to evaluate all three aspects of the pedestrian experience. For example, Noh et al (2020) derived pedestrian trajectories at un-signalized crosswalks through videos from road security cameras and classified crosswalk locations by their safety levels. Traunmueller et al (2018) gathered devices scanned at each Wi-Fi access point (AP), inferred the pedestrian travel trajectories, and identified the most frequently used street segments in Manhattan.…”

Artificial Intelligence Applications to Smart City and Smart Enterprise

“…In reference [9], authors propose a new model for potential pedestrian risky event (PPRE) analysis, using video footage gathered by road security cameras already installed at crossings. The system automatically detects vehicles and pedestrians, calculates trajectories, and extracts frame-level behavioral features.…”