---

Overview

An end-to-end Computer Vision system that detects and counts date palm trees from aerial images and videos using YOLOv11 deep learning model and BoT-SORT multi-object tracking algorithm. Built with a complete production-ready architecture: Flask REST API backend and React Native mobile frontend.

This project addresses the real-world challenge of automated agricultural monitoring, helping farm operators and agricultural researchers count and track date palm trees across large fields without manual effort.

---

Key Features

• YOLOv11 deep learning detection trained on custom aerial imagery dataset
• BoT-SORT multi-object tracking assigns unique IDs to each tree across video frames
• Image and video input support
• Real-time inference via Flask REST API
• Mobile-first React Native frontend
• Tree counting logic eliminates duplicate counts using unique IDs
• Custom dataset prepared and annotated using Roboflow
• Modular architecture with clean separation of concerns

---

Tech Stack

Backend and Machine Learning:

• Python 3.8+
• PyTorch
• YOLOv11 (Ultralytics)
• BoT-SORT
• OpenCV
• Flask

Frontend and DevOps:

• React Native
• Expo
• npm
• Git and GitHub
• Roboflow (Dataset management)

---

System Architecture

React Native Frontend | | HTTP Request (image/video) v Flask REST API | v YOLOv11 Detector (Object Detection) | v BoT-SORT Tracker (Multi-Object Tracking + Counting) | v JSON Response (count, bounding boxes, tracking IDs)

text

---

Project Structure

PalmTreeDetection/ │ ├── backend/ │ ├── detectors/ │ ├── utils/ │ ├── app.py │ ├── config.py │ └── requirements.txt │ ├── frontend/ │ ├── app/ │ ├── components/ │ ├── screens/ │ ├── services/ │ └── package.json │ └── README.md

text

---

Installation and Setup

Prerequisites

• Python 3.8 or higher
• Node.js 16+ and npm
• Expo CLI for React Native
• Git

Backend Setup

git clone https://github.com/MuhammadYasir85a/PalmTreeDetection.git cd PalmTreeDetection/backend

python -m venv venv source venv/bin/activate

pip install -r requirements.txt

python app.py The backend will start on http://localhost:5000

Frontend Setup Bash

cd ../frontend

npm install

npx expo start Scan the QR code with the Expo Go app on your phone (iOS or Android).

How It Works Image/Video Upload — User uploads aerial imagery via mobile app API Request — Frontend sends file via HTTP POST to Flask backend YOLOv11 Detection — Model identifies palm trees with bounding boxes and confidence scores BoT-SORT Tracking — For videos, tracks each tree across frames with unique IDs Counting and Visualization — System counts unique trees and overlays results on output Response Display — Mobile app displays annotated output and total count Use Cases Agricultural monitoring for tracking date palm tree growth Crop yield estimation based on tree count Land surveying and automated counting for large plantations Drone integration for real-time field analysis Research applications supporting agricultural data collection Future Improvements Cloud deployment on Microsoft Azure or AWS Real-time drone video stream support Tree health classification (healthy, diseased, dead) Geo-location tagging using GPS metadata Historical tracking of tree counts over time Multi-class detection for different tree species Web dashboard for analytics and reporting Performance Notes The system demonstrates reliable and consistent performance under standard conditions, with scope for further improvement in complex environments such as dense vegetation overlap, varying lighting conditions, and weather-induced visibility issues.

Author Muhammad Yasir

Computer Science Undergraduate at Namal University Mianwali Aspiring AI and Computer Vision Engineer

Acknowledgments Ultralytics for the YOLOv11 framework Roboflow for dataset annotation tools (Verified Student access) Murtaza's Workshop for OpenCV foundational tutorials Namal University Mianwali for academic guidance License This project is licensed under the MIT License.

```