Autonomous drone vision system using ROS and PyTorch for real-time object classification. Enables drones to make informed decisions based on their surroundings.
- Clone the repo into your ROS workspace
git clone https://github.com/0xVR/DroneAI-Vision-Navigator.git- Build the packages
colcon build- Source the environment to include the newly built packages.
source install/setup.bashIf you are using a different shell (like zsh), adjust the setup script accordingly.
A barebones decision making plugin has been included in the plugins/ directory. This file may be edited to add decision-making functionality to the drone, based on the image classification.
The model is a pretrained ResNet 18, so it's based on the ImageNet categories, which can be found here
Once ready, run the ROS nodes and update dependencies in your existing package.xml and CMakeLists.txt.