Real-Time-Bottle-Detection

🍶 Real-Time-Bottle-Detection using YOLOv8

A custom YOLOv8 object detection model trained to detect and classify water bottles in real-time
Demo • Dataset • Training • Results • Quick Start

---

📌 Overview

This project builds a custom YOLOv8 object detection model trained to detect water bottles and classify them into two categories.

Class	Icon	Description
plastic_bottle	🧴	Transparent or colored plastic water bottles
steel_bottle	🥤	Metal or stainless steel water bottles

The model draws bounding boxes with confidence scores around detected bottles in any image or video feed.

---

📸 Demo

After Detection

✅ Model detected plastic_bottle and steel_bottle with 0.89 confidence

---

🛠️ Tech Stack

Tool	Purpose
Python 3.8+	Programming language
YOLOv8 (Ultralytics)	Object detection model
Google Colab	Cloud GPU training
Roboflow	Dataset preparation and annotation
OpenCV	Image processing
PyTorch	Deep learning framework
Matplotlib	Visualization and charts

---

📊 Dataset

Detail	Info
Source	Roboflow Universe
Total Images	315
Training Set	250 images
Validation Set	26 images
Test Set	39 images
Annotation Format	YOLOv8 Bounding Boxes
Input Size	512×512
Classes	plastic_bottle, steel_bottle

Dataset link:

dataset/dataset_link.txt

---

🏋️ Training

Training was performed on Google Colab using a Tesla T4 GPU.

Training Configuration

from ultralytics import YOLO

model = YOLO("yolov8n.pt")

model.train(
    data="data.yaml",
    epochs=50,
    imgsz=640
)

Training Parameters

Parameter	Value
Model	YOLOv8 Nano
Epochs	53
Image Size	640×640
GPU	Tesla T4
Framework	Ultralytics

---

📈 Results

Performance Metrics

Metric	Score
mAP@50	0.806
mAP@50-95	0.618
Precision	0.913
Recall	0.804
Inference	43.8 ms

---

📊 Training Charts

---

🚀 Quick Start

1. Clone Repository

git clone https://github.com/JonnadaRaju/Real-Time-Bottle-Detection.git
cd Real-Time-Bottle-Detection

2. Install Dependencies

pip install ultralytics opencv-python matplotlib

3. Run Detection on Image

from ultralytics import YOLO

model = YOLO("models/best.pt")

results = model.predict(
    source="images/demo/input_plastic.jpg",
    save=True,
    conf=0.5
)

print(f"Detected: {len(results[0].boxes)} bottles")

4. Run Webcam Detection

from ultralytics import YOLO

model = YOLO("models/best.pt")

model.predict(
    source=0,
    show=True,
    conf=0.5
)

5. Run Video Detection

from ultralytics import YOLO

model = YOLO("models/best.pt")

model.predict(
    source="your_video.mp4",
    save=True,
    conf=0.5
)

---

🧠 Skills Demonstrated

Computer Vision
Object Detection
YOLOv8
Deep Learning
Dataset Annotation
Model Training
Model Inference
Python
Google Colab
Roboflow
PyTorch
OpenCV
Transfer Learning

---

Made with ❤️ using YOLOv8 + Google Colab ⭐ Star this repo if you found it helpful!