Detection Dataset
2970
Total images with manually labeled bounding boxes.
Overview
Existing datasets do not capture the diversity of vehicle types and non-lane-disciplined traffic conditions in Sri Lanka. SL-LPR addresses this gap with carefully collected and annotated real-world traffic data.
- Multiple vehicle types: cars, buses, trucks, motorbikes, three-wheelers
- Multi-lane scenes with varying traffic densities
- Daytime and low-light (dusk) conditions
Dataset Highlights
Recognition Dataset
3412
License plates standardized to the CAA 0923 8-character format.
Recognition Quality
Duplicate filtering applied (max 3 per plate with varying clarity) and unreadable images removed.
Sample Data
Examples from the dataset are shown below.
Team
The SL-LPR dataset was developed by the project team with academic supervision and industry support.
Research Applications
- License Plate Detection (LPD)
- License Plate Recognition (LPR)
- Intelligent Transportation Systems (ITS)
Associated Paper
Title: An Embedded Real-Time License Plate Recognition System for Complex Traffic Scenes
Accepted at: IEEE Intelligent Transportation Systems Society Conference (ITSC 2026), Naples, Italy
Project explanation: Watch on YouTube
Paper link: Will be added soon
Citation: Will be added soon
How to Request Access
Click the request form, complete all fields, and tick the agreement checkbox before submitting.
Email: manimohant.20@uom.lk
User agreement: Read agreement page
Request form: Open dataset access form
Prefer using a university/institution email. Typical response time is 3-7 working days.
Usage Policy
- Dataset is for research purposes only.
- No redistribution allowed.
- No commercial use.
- Must cite the paper once available.
Updates
| Date | Update |
|---|---|
| 2026 | Initial release |
| Upcoming | Paper and citation details |
Acknowledgment
We are grateful to the following parties who helped make this project successful.
Dr. Sampath Perera and Eng. Kithsiri Samarasinghe actively supervised the project and provided guidance. Dr. Ajith Pasqual connected us with this project and important resource personnel. The staff of PEX Microsystems informed us of industry standards, provided professional opinions, and assisted with development board selection. PEX Microsystems also provided funding for the development boards.
Prof. Saman Bandara, Member of the National Council for Road Safety (NCRS), helped to submit a proposal to the NCRS.
Dr. Ranga Rodrigo and the National Research Council of Sri Lanka (NRC, Grant No. 19-080) provided us with the computational resources needed for training the machine learning models.
Further, we are grateful to the Sri Lanka Police and the Road Development Authority (RDA) for giving us permission to test our product on normal roads and the Southern Expressway.
We truly appreciate their combined expertise, mentorship, and support throughout this project.