Automating Crash Diagram Generation Using Vision-Language Models: A Case Study on Multi-Lane Roundabouts

TL;DR

This paper explores automating crash diagram creation from police reports using Vision-Language Models, demonstrating promising results with GPT-4o and highlighting current limitations in engineering visualization tasks.

Contribution

It introduces a structured prompt framework for VLMs to generate crash diagrams and evaluates model performance on multilane roundabout crash reports.

Findings

GPT-4o achieved the highest performance score of 6.29/10.

VLMs show promise but still face limitations in engineering visualization.

The study provides a foundation for integrating AI into crash analysis workflows.

Abstract

Crash diagrams are essential tools in transportation safety analysis, yet their manual preparation remains time-consuming and prone to human variability. This study investigates the use of Vision-Language Models (VLMs) to automate crash diagram generation from police crash reports, focusing on multilane roundabouts as a challenging test case. A three-part structured prompt framework was developed to guide model reasoning through interpretation, extraction, and visual synthesis, while a 10-metric evaluation system was designed to assess diagram quality in terms of semantic accuracy, spatial fidelity, and visual clarity. Three popular models, including GPT-4o, Gemini-1.5-Flash, and Janus-4o, were tested on 79 crash reports. GPT-4o achieved the highest average performance (6.29 out of 10), followed by Gemini-1.5-Flash (5.28) and Janus-4o (3.64). The analysis revealed GPT-4o's superior spatial reasoning and alignment between extracted and visualized crash data. These results highlight both the promise and current limitations of VLMs in engineering visualization tasks. The study lays the groundwork for integrating generative AI into crash analysis workflows to improve efficiency, consistency, and interpretability.