Pixels to Signals: A Real-Time Framework for Traffic Demand Estimation

2025-10-28

AI Summary

To address the challenges of real-time traffic demand estimation and high deployment costs in urban congestion management, this paper proposes a lightweight pixel-sequence analysis method for efficient vehicle detection. The approach extracts motion foreground via temporal background modeling from video streams and employs DBSCAN density-based clustering for accurate vehicle localization and counting—eliminating reliance on complex deep learning models or hardware modifications. Compared to conventional methods, it offers low computational overhead, flexible deployment, and strong scalability, making it suitable for real-time traffic perception across large-scale urban road networks. Experimental results demonstrate high detection accuracy (mAP@0.5 > 92%) with sub-30 ms per-frame processing time on 1080p video, significantly lowering the barrier for edge-device deployment. This enables reliable, cost-effective data acquisition for downstream applications such as traffic flow prediction and adaptive signal control.

Technology Category

• Computer Vision: Motion & Tracking
• Machine Learning: Learning on the Edge & Model Compression
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Experiences and lessons learnt from Web-based algorithms and system deployments
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Traffic congestion is becoming a challenge in the rapidly growing urban cities, resulting in increasing delays and inefficiencies within urban transportation systems. To address this issue a comprehensive methodology is designed to optimize traffic flow and minimize delays. The framework is structured with three primary components: (a) vehicle detection, (b) traffic prediction, and (c) traffic signal optimization. This paper presents the first component, vehicle detection. The methodology involves analyzing multiple sequential frames from a camera feed to compute the background, i.e. the underlying roadway, by averaging pixel values over time. The computed background is then utilized to extract the foreground, where the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm is applied to detect vehicles. With its computational efficiency and minimal infrastructure modification requirements, the proposed methodology offers a practical and scalable solution for real-world deployment.

Problem

• Estimating real-time urban traffic demand using camera feeds
• Detecting vehicles through background modeling and DBSCAN clustering
• Optimizing traffic flow to reduce congestion and delays

Innovation

• Real-time vehicle detection using sequential camera frames
• Background modeling through pixel value averaging
• DBSCAN clustering algorithm for foreground vehicle identification