Case study- Smart City Project

Automatic License Plate Detection Solution - Smart City Project

Summary

An industry leader in comprehensive video content analytics platforms based in East Asia sought to enhance its surveillance services to make cities smarter and safer. They approached Folio3 AI to develop an Automatic Number Plate Recognition (ANPR) solution capable of reading vehicle number plates from images, videos, and live camera streams. The solution needed to handle both printed and handwritten number plates in a standard format approved by the authorities and be deployable on-site.

About the Customer

The customer is a prominent company in East Asia specializing in Video Content Analytics Platforms. Their mission is to enhance urban safety and intelligence through advanced surveillance technologies. With a significant market presence and a commitment to innovation, they continuously seek to improve their solutions to meet evolving security needs.

  • Team composition

    4 members

  • Expertise used

    Machine Learning, Computer Vision, OCR, ALPR

  • Duration

    4 months

  • Services provided

    Model training, Data capture, On-Site Deployment

  • Region

    Southeast Asia

  • Industry

    Cyber Intelligence, and Surveillance

Understanding the Challenge

The client required a robust and accurate solution to automatically read vehicle number plates from various input sources, including live camera feeds, images, and videos. The challenge was to develop a system capable of recognizing both printed and handwritten number plates in the standard format approved by authorities. Additionally, the solution needed to be deployable on-site to integrate seamlessly with their existing surveillance infrastructure.

Solution

Folio3 AI developed a comprehensive ANPR solution tailored to the client's requirements.

Input Handling

The application could take input from live camera feeds, images, and videos.

input handling

Frontend Application

A user-friendly frontend application was developed, allowing users to upload images, and videos, or input live camera feeds.

frontend application

High Accuracy

Leveraging advanced machine learning algorithms, the solution provided high accuracy in number plate recognition.

high accuracy (1)

User-Friendly Interface

The frontend application allowed users to easily upload and manage input sources.

user friendly interface
object blur-second section
ai lisence plate reading

AI License plate reading
The system was designed to accurately read both printed and handwritten number plates in the standard format.

on site deployement

On-Site Deployment
The solution was designed for on-site deployment, ensuring smooth integration with the client's existing surveillance systems.

versality

Versatility
Capable of processing various input formats, including live camera feeds, static images, and recorded videos.

robustness

Robustness
The solution was designed to handle diverse environmental conditions and variations in number plate designs.

Result

With an accuracy above 90%, the successful on-site deployment ensured seamless integration with the client's existing systems, contributing to their mission of making cities smarter and safer.