Abstract:Accurate and reliable tracking of multi-pedestrian is of great importance for autonomous driving, human-robot interaction and video surveillance. Since different scenarios have different best-performing sensors, sensor fusion perception plans are believed to have complementary modalities and be capable of handling situations which are challenging for single sensor. In this paper, we propose a novel track-to-track fusion strategy for multi-pedestrian tracking by using a millimeter-wave (MMW) radar and a monocular camera. Pedestrians are firstly tracked by each sensor according to the sensor characteristic. Specifically, the 3D monocular pedestrian detections are obtained by a convolutional neural network (CNN). The trajectory is formed by the tracking-by-detection approach, combined with Bayesian estimation. The measurement noise of the 3D monocular detection is modeled by a detection uncertainty value obtained from the same CNN, as an approach to estimate the pedestrian state more accurately. The MMW radar utilizes the track-before-detection method due to the sparseness of the radar features. Afterwards, the pedestrian trajectories are obtained by the proposed track-to-track fusion strategy, which can work adaptively under challenging weather conditions, low-illumination conditions and clutter scenarios. A group of tests are carried out to validate our pedestrian tracking strategy. Tracking trajectories and optimal sub-pattern assignment (OSPA) metric demonstrate the accuracy and robustness of the proposed multi-sensor multi-pedestrian tracking system.

Study on Pedestrian Detection and Tracking with Monocular Vision

A Novel Approach to Design the Fast Pedestrian Detection for Video Surveillance System

Online Multipedestrian Tracking Based on Fused Detections of Millimeter Wave Radar and Vision

The study of the detection and tracking of moving pedestrian using monocular-vision

A Monocular Vision Based Pedestrian Detection System for Intelligent Vehicles

Algorithm Study for Pedestrian Detection Based on Monocular Vision

Method of moving pedestrian detection and tracking based on monocular vision technology

Research Of Pedestrian Detection For Intelligent Vehicle Based On Machine Vision

Research on Pedestrian Object Detection and Trajectory Tracking Methods

Adaptive Multi-Pedestrian Tracking by Multi-Sensor: Track-to-Track Fusion Using Monocular 3D Detection and MMW Radar

Pedestrian Detection and Tracking Based on Automotive Vision

Research on Vision-based Pedestrian Detection and Tracking Algorithm

A Real-Time Active Pedestrian Tracking System Inspired by the Human Visual System

Effective multiple pedestrian tracking system in video surveillance with monocular stationary camera

Study on Pedestrian Detection Ahead of Vehicle Based on Machine Vision

Study on Pedestrian Tracking Based on Kalman Filter

Online Adaptive Multiple Pedestrian Tracking in Monocular Surveillance Video

Monocular pedestrian tracking from a moving vehicle

Pedestrian Detection and Tracking Using HOG and Oriented-Lbp Features

Pedestrian Tracking Based on Camshift with Kalman Prediction for Autonomous Vehicles

Vision Based Multi-pedestrian Tracking Using Adaptive Detection and Clustering.