Visual control of the Parrot drone with OpenCV, Ros and Gazebo Simulator

Abstract

Drones are becoming more and more popular these days. There are multiple tasks and ideas that can be implemented to them. One of them is the idea described in that work. The aim of the project was creating a software in C++ in ROS that will control the Parrot Drone simulated in Gazebo Simulator. The drone movements are stimulated by the orange ball movements in front of the camera (the picture below). As it is presented on the diagram below, there were needed: Ardrone Autonomy Package (drone driver), Gazebo Simulator, Tum Simulator (connection of two previous ones), and cv_bridge in OpenCV library, which converts ros image messages to OpenCV objects. Then the objects can be modified in OpenCV library. The evaluation of project had two main phases. First of them was to configure Ardrone Autonomy driver and Gazebo Simulator. Second step of that phase was to connect two previous ones by configuring TUM Simulator. Second main phase was to write a program that converts ros image messages frome the camera into openCV images, analyzes them and depending on the movement of the ball controlls the drone movement. The target was achieved and the drone is being controlled by the ball in front of the camera.