Related Experiment Video
Updated: Jan 1, 2026
16:14
Trajectory Data Analyses for Pedestrian Space-time Activity Study
Published on: February 25, 2013
14.0K
An Approach to Segment and Track-Based Pedestrian Detection from Four-Layer Laser Scanner Data
Mingfang Zhang1, Rui Fu2, Wendong Cheng2
1Beijing Key Lab of Urban Intelligent Traffic Control Technology, North China University of Technology, Beijing 100144, China.
Sensors (Basel, Switzerland)
|December 15, 2019
Summary
This study introduces a new pedestrian detection method using a four-layer laser scanner, improving accuracy for occluded pedestrians by integrating tracking data with segment classification.
Area of Science:
- Robotics
- Computer Vision
- Artificial Intelligence
Background:
- Pedestrian detection is crucial for autonomous driving systems.
- Challenges include variations in human appearance, pose, and partial occlusion.
Purpose of the Study:
- To develop a novel pedestrian detection method using a four-layer laser scanner.
- To address the challenge of partial occlusion by integrating tracking information.
Main Methods:
- Point cloud segmentation into object clusters.
- Extraction of 18 effective features using selection algorithms and correlation analysis.
- Track classification using particle filter and probability data association filter.
Main Results:
- Both back-propagation neural networks and Adaboost classifiers showed advantages in segment classification.
- Track classification significantly improved detection of partially occluded pedestrians.
Conclusions:
- The proposed method enhances pedestrian detection reliability, especially in challenging occluded scenarios.
- Fusion of segment classification and tracking knowledge offers a robust solution for intelligent vehicles.
