Automated Analysis of Road Safety with Video Data

Abstract

Traffic safety analysis has often been undertaken with historical collision data. However, well-recognized availability and quality problems are associated with collision data. In addition, the use of collision records for safety analysis is reactive: a significant number of collisions has to be recorded before action is taken. Therefore, the observation of traffic conflicts has been advocated as a complementary approach in the analysis of traffic safety. However, incomplete conceptualization and the cost of training observers and collecting conflict data have been factors inhibiting extensive application of the traffic conflict technique. The goal of this research is to develop a method for automated analysis of road safety with video sensors to address the problem of dependency on the deteriorating collision data. The method automates the extraction of traffic conflicts from video sensor data. This method should address the main shortcomings of the traffic conflict technique. A comprehensive system is described for traffic conflict detection in video data. The system is composed of a feature-based vehicle tracking algorithm adapted for intersections and a traffic conflict detection method based on the clustering of vehicle trajectories. The clustering uses a K-means approach with hidden Markov models and a simple heuristic to find the number of clusters automatically. Traffic conflicts can then be detected by identifying and adapting pairs of models of conflicting trajectories. The technique is demonstrated on real-world video sequences of traffic conflicts.

Analysis of Pedestrian–Vehicle Traffic Conflicts in Street Designs with Elements of Shared Space

“…While initial methods were purely qualitative and used only subjective statements to identify and characterise evasive actions of road users in response to conflicts, newer methods have become more quantitative and have provided definitions for measures of the severity of conflicts (18)(19)(20). This has also been facilitated by advances in imaging technology, which now enables the automated observation of sites and application of traffic conflicts techniques is a much simpler way, as opposed to earlier techniques based on manual observation and the use of grading sheets (21)(22). Notable examples of techniques in use today include: the "Swedish Traffic Conflicts Technique" (STCT) from Lund University in Sweden (23), which focuses upon evaluating the time to accident and vehicle speed at the beginning of an evasive action using a graph, which has disaggregating lines to categorise the different conflicts according to severity; the "CIHT Conflicts Technique" (CIHTCT) from the Transport and Road Research Laboratory in the UK (14), which uses a categorisation of the various comprising elements of the conflict and creates levels of severity for each element, such that by summing up the given element levels an overall grading of the severity of the conflict can be obtained; and the "USDoT Conflict Technique" (USDoTCT) from the FHWA in the US (13), which is similar in concept to the CIHTCT but uses different grading criteria.…”

Evaluating Safety at Railway Level Crossings with Microsimulation Modeling

“…Thereafter, the internationally accepted definition of a traffic conflict became (and remains) "an observable situation in which two or more road users approach each other in space and time to such an extent that there is a risk of collision if their movements remain unchanged" (12). This alternative definition has been adopted in a number of safety studies (13)(14)(15)(16)(17). Sayed and Zein developed predictive models that use linear regression analysis to relate the number…”

Probabilistic Framework for Automated Analysis of Exposure to Road Collisions

“…In a previous paper (16), the authors have shown that traffic conflicts can be detected in video sequences using HMM-based semisupervised machine learning techniques. This paper presents an extension of the earlier work (16,17): a comprehensive probabilistic framework relying on the concept of the safety hierarchy. It provides a computational definition of severity as the probability of collision that is suited for an automated system.…”