🍎 Fruit Detection (YOLOv8)

Real-time fruit detection powered by YOLOv8, using a custom fruit dataset split into train/val/test.
Built for quick deployment and clean reproducibility.

---

🚀 Features

• 🧠 YOLOv8n model trained on a custom fruit dataset from Roboflow
• 🧩 Custom Split: dataset manually divided into train, val, and test sets
• 🎯 Pretrained Weights: runs directly using best.pt without retraining
• 💻 Simple Interface: just run one Python file - no complex setup needed

---

📄 Essay Report

The full essay report has been uploaded to Google Drive.
You can access it here 👉
🔗 Download Essay Report

---

🍉 Dataset Overview

The dataset contains 6 fruit classes used for object detection:

Class	Description
🍍 Pineapple	Tropical fruit with spiky skin and sweet yellow flesh.
🍒 Cherry	Small red fruit often appearing in pairs.
🥭 Mango	Yellow-orange fruit with smooth skin and sweet aroma.
🍑 Plum	Round fruit with smooth skin, purple or red when ripe.
🍅 Tomato	Red juicy fruit often mistaken for a vegetable.
🍉 Watermelon	Large green fruit with red interior and black seeds.

---

🗂 Project Structure

Fruits-Detection/
├── program.py               # Main entry point for running detection
├── requirements.txt         # Python dependencies
├── LICENSE
├── SECURITY.md
│
├── weights/
│   ├── best.pt              # Trained YOLOv8 model weights
│   └── last.pt
│
└── dataset_fruits/
    ├── data.yaml            # Dataset configuration for YOLOv8
    ├── README.dataset.md    # Info on dataset source & how val set was created
    ├── README.roboflow.txt  # Original Roboflow export metadata (source information)
    ├── train/               # Training images & labels
    ├── valid/               # Validation images & labels
    └── test/                # Test images & labels

---

⚙️ Installation

1️⃣ Clone the repository

git clone https://github.com/quan831/Fruit-Detection.git
cd Fruit-Detection

2️⃣ (Optional) Create virtual environment

python -m venv venv
# Activate:
# Windows
venv\Scripts\activate
# macOS / Linux
source venv/bin/activate

3️⃣ Install dependencies

pip install -r requirements.txt

---

▶️ Run Detection

🧩 Option 1 — Detect via Script

python program.py

Make sure your working directory includes:

• weights/best.pt
• dataset_fruits/data.yaml

The program loads the YOLOv8 model and runs inference directly.

🧠 Option 2 — Run in Spyder (Recommended for GUI)

1. Open Anaconda Navigator → Launch Spyder
2. Open program.py
3. Hit Run (F5) to start detection
4. Check outputs and logs inside the console or generated output folder (if any)

---

🧠 Model Details

• Model: best.pt (trained YOLOv8n)
• Framework: Ultralytics YOLOv8 (Python)
• Dataset: Custom split version of Roboflow fruit dataset
• Train/Val/Test Ratio: defined manually in README.dataset.md

---

📸 Preview

• Image Detection Interface:
• Real-time Camera Detection Interface:

---

🧾 License

This project is licensed under the MIT License.

---

🛡 Security

See SECURITY.md for details on responsible disclosure.

---

💖 Credits

• Ultralytics YOLOv8
• Roboflow for dataset hosting
• Original dataset: nhận diện trái cây v2 Computer Vision Dataset
• Custom dataset split and model tuning by Quan (James)

---

🌟 Show Your Support

If this repo helps you, please give it a ⭐ on GitHub - it really motivates me!

---

👤 Authors

• Nguyễn Minh Quân (Leader)
• Hoàng Quốc Khánh
• Lê Hoàng Lan
• Triệu Yến Vi