EvilNET flying objects detector

This is streamlit project with functionality to monitor areas close to airports.

Supported classes:

• copter
• plane
• helicopter
• bird
• drone (plane-type)

Used technologies

• Streamlit
• YOLO
• Docker
• SQLite

Project requirements

• Nvidia GPU + CUDA
• Python >=3.9
• Windows 10/11 / Linux / Docker

Functionality

• Setting up video streaming devices (cameras)
• Single image detection
• Video detection
• Multiple images detection (return .zip with .txt)
• Configurable confidence threshold and image size
• Download / delete video in database-like view

Installation

via Docker

Prerequisites:

• Linux / WSL host system
• GPU + drivers
• Allow docker containers to use host system's GPUs

git clone https://github.com/EvilSumrak2049/EvilNET.git
cd EvilNET
docker build -t streamlit .
docker run -p 127.0.0.1:5443:8501 --gpus all streamlit

manually (Windows / Linux)

Follow these instructions:

git clone https://github.com/EvilSumrak2049/EvilNET.git
cd EvilNET
pip install torch torchvision --index-url https://download.pytorch.org/whl/cu121
pip install -r requirements.txt
streamlit run design.py --server.maxUploadSize=100000