Icarus Eye is built using a custom YOLOv8s model trained on the VisDrone2019-DET Dataset. This project is a semester project for a Machine Vision course at KSU. The project aims to be a proof of concept for real-time object detection and tracking using a drone. The project is built using Python, OpenCV, and PyTorch. The project is built to be modular and customizable, allowing for easy integration with other models and datasets. The project's name is a pun and a reference to the Greek myth of Icarus, who flew too close to the sun and fell, and the vigilant Eye of Sauron from the Lord of the Rings.
- Description
- Introduction
- Features
- Installation
- Demonstration / Examples
- Usage
- Limitations
- Future Work
- License
The project is built using a custom YOLOv8s model trained on the VisDrone2019-DET Dataset. The project is built with Python utilizing the following libraries:
| 📚 Library | 🗒️ Description | Version | 🌐 Website | 🥧 PyPi Repo |
|---|---|---|---|---|
| Ultralytics | Open Source Object Detection Library | 8.3.1 | Ultralytics.com | ultralytics |
| OpenCV | Open Source Computer Vision Library | 4.10.0.84 | OpenCV.org | opencv-python |
| PyTorch | Open Source Machine Learning Library | 2.4.1 | PyTorch.org | torch |
| PyQt6 | Python Binding for Qt | 6.7.1 | RiverbankComp | PyQt6 |
| Numpy | Scientific Computing Library | 1.26.4 | NumPy.org | numpy |
| MatPlotLib | Data Visualization Library | 3.5.1 | MatPlotLib | matplotlib |
The core of the project is built using the Ultralytics YOLOv8s model along with Ultralytics ByteTrack implementation.
The project was built with the edge device in mind, and as such, the project is optimized for performance / tuning. The project has the following features:
- Easy swapping of Ultralytics YOLO models.
- Quick and easy modification of output tuning.
- Performance optimizations
- Max-Bounding Box (Sorted by Confidence)
- Nth Frame Detection / Tracking
- Framerate control on pre-recorded video
- Live list of omitted classes
- Toggleable tracking
- and more...
Regardless of the installation method, the project heavily benefits from ffmpeg being installed on the system. This
allows for the video device detection, settings, and controls to be more accessably and user-friendly.
The following
link will guide you through the installation process for ffmpeg on your system:
FFmpeg Installation
The following command will install ffmpeg on your system:
sudo apt-get install ffmpegThe following command will install ffmpeg on your system utilizing brew:
brew install ffmpegWhen installing the project, it is recommended to use a virtual environment. Along with this, since the project
utilizes torch and torchvision, it is recommended to install the libraries with CUDA support. Without this, the
project will run on the CPU, which will be significantly slower. The following steps will guide you through the
installation process.
- (Optional) Create a virtual environment:
python -m venv venv
source venv/bin/activate # Linux / MacOS
venv\Scripts\activate # Windows- Clone the repository:
git clone https://github.com/JonathanLTurner03/IcarusEye.git
cd IcarusEyePrior to this next step, if you have a CUDA compatible GPU, you can modify the requirements.txt file to include the
appropriate version of torch and torchvision. If you do not have a CUDA compatible GPU, you can skip this step.
- Install the required libraries:
pip install -r requirements.txt - Run the project:
python3 main.pyIn Progress
In Progress
The project is built to be user-friendly and easy to use. The project is built with a GUI using PyQt6. The following steps will guide you through the usage of the project.
Changing the Model
- Click the
Modeldropdown menu. - Select change YOLO model.
- Select the model you wish to use.
- If the model is not in the dropdown or you wish to use a custom model:
- click the
Customoption and select the model's.ptfile. - Click
OK.
- click the
- Click
Submitto apply the changes.
Loading a pre-recorded video
- Click the
File Inputradio button. - Click the
Select Video Filebutton. - Find and select the video file you wish to use utilizing the popup
explorerfile selector. - After selecting the video file, click
Playto start the video playback.
Using a webcam / capture device
- Click the
Device Inputradio button (if not already selected as default). - Click the
Select Devicedropdown menu and select the capture device wanted. - If wanting to use the device's self-reported settings, select the
Automaticradio button and thenSubmit. NOT RECOMMENDED - If wanting to use custom settings, select the
Manualradio button.- Select the
FPSnumber field and specify the desired FPS. - If needed, change the codec to a fourcc codec.
- Select the
ResolutionTextbox and specify the desired resolution *Note: it must be in the format ofwidthxheight. - Click
OK.
- Select the
- After selecting the device and settings, click
Playto start the video feed.
Still In Progress
In Progress
In Progress
Licensed under the MIT License. See LICENSE for more information.
MIT License
Copyright (c) 2024 Jonathan Turner
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.