English日本語 • 
login

Pedestrian Trajectory Mapping with MapLibre and OpenCV from Smartphone Videos
.ical
2026-09-02 –, Ran2

I needed to analyze pedestrian flow patterns and draw them on map in urban spaces, so I built a simple pipeline to extract trajectories from smartphone videos which shot from a low line of sight.

Overall Pipeline

  1. Capturing videos of pedestrians with smartphone
  2. Pedestrian Detection with openCV DNN and makeing trajectories with CentroidTracker
  3. Homography transformation from video coordinate to wgs84
  4. Draw trajectory on interactive map

In this session

I will talk about the barriers encountered during the experiment.
1. From holding the smartphone with my hand to securing it with a tripod.
2. How effective is a low line of sight in videos.
3. Comparison between centroid and footprint trackers.
4. How and how many trajectory points I sampled.

Technologies Used

This project utilizes only open source technologies:
- OpenCV DNN - Computer vision and deep learning
- CentroidTracker - Multi-object tracking
- MapLibre GL - Interactive web mapping
- YOLOx - Object detection
- Python, TypeScript, React - Development frameworks

License

This project acknowledges the following open source licenses:
- Apache 2.0 - OpenCV, YOLOx, TypeScript
- BSD 3-Clause - MapLibre GL
- MIT - React
- Python Software Foundation License - Python

I expect that this project contributes to individual protest analysis by analyzing pedestrian reactions.

Level of technical complexity: 2 - intermediate Give indication of resources (video, web pages, papers, etc.) to read in advance, that will help get up to speed on advanced topics.:

https://github.com/TOKIHISA/people_trajectory_analysis

Indicate what is (are) the open source project(s) essential in your talk: I make my conference contribution available under the CC BY 4.0 license. The conference contribution comprises the abstract, the text contribution for the conference proceedings, the presentation materials as well as the video recording and live transmission of the presentation:
The speaker’s profile picture
Toki Hirose

I am a geospatial data analyst with 5+ years of experience in analyzing spatial phenomena.
I aspire to develop measurement and analysis capabilities that span from natural environments to complex social dynamics, using open source geospatial tools.