Spaces:
Sleeping
Sleeping
File size: 2,676 Bytes
eccf2ca 2d21e63 5f18f57 607d20a eccf2ca f7c1359 2d21e63 5f18f57 2d21e63 5f18f57 2d21e63 5f18f57 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 | ---
title: PlateVision
emoji: 🚗
colorFrom: indigo
colorTo: pink
sdk: streamlit
pinned: false
sdk_version: 1.50.0
short_description: AI-Powered License Plate Detection & Recognition from Images
---
# 🚗 PlateVision 🔍
### AI-Powered License Plate Detection & Recognition from Images and Videos
---
## 🧠 Overview
**PlateVision** is an AI-powered web app built with **Streamlit** that detects and recognizes vehicle license plates from **images** and **videos** using a **fine-tuned YOLOv8 model** and **EasyOCR**.
It seamlessly combines **object detection** and **optical character recognition (OCR)** to deliver a simple, intuitive, and interactive platform where users can upload their media files and get annotated results in seconds — complete with **confidence scores** and **recognized text overlays**.
---
## ✨ Features
✅ Detects license plates from both **images** and **videos**
✅ Extracts **license numbers** using OCR (EasyOCR)
✅ Displays **detection confidence** and recognized text on output
✅ Supports multiple formats — `.jpg`, `.jpeg`, `.png`, `.mp4`, `.mkv`
✅ Shows **real-time progress** while processing videos
✅ Provides a **clean Streamlit interface** for easy interaction
✅ Outputs processed results with bounding boxes and plate numbers drawn on frames
---
## ⚙️ How It Works
1. **Upload Input File**
- The user uploads an image or video through the Streamlit interface.
2. **License Plate Detection (YOLOv8)**
- A **custom fine-tuned YOLOv8 model** detects the position of license plates in the media.
3. **Text Recognition (EasyOCR)**
- The detected plate region is passed through **EasyOCR** to read and extract the license number.
4. **Annotation & Output**
- Detected plates are annotated with bounding boxes, confidence percentages, and extracted text.
- For videos, frames are processed sequentially and stitched back together using **MoviePy**.
5. **Display**
- The processed media (image or video) is displayed directly in the Streamlit app.
---
## 🧩 Tech Stack
| Component | Technology / Library | Purpose |
|------------|----------------------|----------|
| 🧱 Frontend & UI | **Streamlit** | Builds the interactive web interface |
| 🧠 Object Detection | **YOLOv8 (Ultralytics)** | Detects license plate locations |
| 🔍 OCR Engine | **EasyOCR** | Recognizes alphanumeric text from plates |
| 🎥 Video Handling | **MoviePy** | Processes and reconstructs video frames |
| 👁️ Computer Vision | **OpenCV** | Image manipulation and annotation |
| 💾 Language | **Python 3.11+** | Development language |
---
|