Dynamic Hackathon Project - Team Arjun (Team Number 348)

Team Members

• Viren Mehta
• Karan Patel
• Perin Modi
• Shrey Patel

---

Introduction

This project is designed for collision avoidance in multi-crane environments. The system utilizes machine learning models to detect and track persons near crane jibs to prevent accidents. The implementation includes dataset creation, training models, and a GUI-based simulation for real-time detection and tracking.

The model trained is YOLOv8, achieving an accuracy of 96%.

---

How to Use the Code

Follow these steps to use and run the project:

1. Setup Environment

Ensure you have Python installed along with necessary dependencies. Install the required packages using:

pip install -r requirements.txt

2. Dataset Creation

Run the dataset_creation.ipynb notebook to generate and preprocess the dataset for training.

3. Train the Model

Use jib-person-training.ipynb to train the model for detecting people near crane jibs.

4. Person Detection & Tracking

• Run person_detect.ipynb to detect people in the crane operation area.
• Use person_tracker.ipynb to track their movement.

5. Collision Avoidance Simulation

Run the collision_avoidance_multicrane.py script to visualize crane operations, restricted zones, and collision warnings using a Tkinter-based GUI.

python collision_avoidance_multicrane.py

6. Video Playback Prototype

Use prototype_video_gui.py to play prototype demonstration videos in a Tkinter-based player.

python prototype_video_gui.py

---

Project Resources

• PPT: View Here
• Demo Video: Watch Here

---

Contact

• Viren Mehta: LinkedIn
• Karan Patel: LinkedIn
• Perin Modi: LinkedIn
• Shrey Patel: LinkedIn

---

Developed for Dynamic Hackathon 2025 by Team Arjun (348).