Abstract:This chapter presents an approach to detect and track multiple people robustly in real time using laser scan frames. The detection and tracking of people in real time is a problem that arises in a variety of different contexts. Examples include intelligent Surveillance for security purposes, scene analysis for service robot, and crowd behavior analysis for human behavior study. Over the last several years, an increasing number of laser-based people-tracking systems have been developed in both mobile robotics platforms and fixed platforms using one or Multiple laser scanners. It has been proved that processing on laser scanner data makes the tracker much faster and more robust than a vision-only based one in complex situations. In this chapter, we present a novel robust tracker to detect and track multiple people in a crowded and open area in real time. First, raw data are obtained that measures two legs for each people at a height of 16 cm from horizontal ground with multiple registered laser scanners. A stable feature is extracted using accumulated distribution of successive laser frames. In this way, the noise that generates split and merged measurements is smoothed well, and the pattern of rhythmic swinging legs is utilized to extract each leg. Second, a probabilistic tracking model is presented, and then a sequential inference process using a Bayesian rule is described. A sequential inference process is difficult to compute analytically, so two strategies are presented to simplify the computation. In the case of independent tracking, the Kalman filter is used with a more efficient measurement likelihood model based on a region coherency property. Finally, to deal with trajectory fragments we present a concise approach to fuse just a little visual information from synchronized video camera to laser data. Evaluation with real data shows that the proposed method is robust and effective. It achieves a significant improvement compared with existing laser-based trackers.

Laser-based detection and tracking moving objects using data-driven Markov chain Monte Carlo

A Compensatory Algorithm for High-Speed Visual Object Tracking Based on Markov Chain

Probabilistic 3D Multi-Modal, Multi-Object Tracking for Autonomous Driving

Real-Time Detection And Tracking Of Multiple People In Laser Scan Frames

Boosted Markov Chain Monte Carlo Data Association for Multiple Target Detection and Tracking

Vision-Based Moving Objects Detection with Background Modeling

Fusion of Detection and Matching Based Approaches for Laser Based Multiple People Tracking

Moving Object Classification Using Horizontal Laser Scan Data

Multi-modal Tracking of People Using Laser Scanners and Video Camera

Front-view Vehicle Detection by Markov Chain Monte Carlo Method

Multiple-Target Tracking by Spatiotemporal Monte Carlo Markov Chain Data Association

Dynamic Multi-LiDAR Based Multiple Object Detection and Tracking

Real-Time Avoidance Strategy of Dynamic Obstacles via Half Model-Free Detection and Tracking With 2D Lidar for Mobile Robots

A Laser-Scanner-Based Approach Toward Driving Safety and Traffic Data Collection

Bayesian Fusion of Laser and Vision for Multiple People Detection and Tracking

Multi-Object Tracking based on Imaging Radar 3D Object Detection

Detection and Tracking of Moving Objects at Intersections Using a Network of Laser Scanners

Generic Tracking and Probabilistic Prediction Framework and Its Application in Autonomous Driving

Multi-object Detection, Tracking and Prediction in Rugged Dynamic Environments

LADAR-Based Mover Detection from Moving Vehicles

Laser-based Tracking of Multiple Interacting Pedestrians Via On-Line Learning.