Write an Object Tracking Drone Application

摘要

In this article, I discuss how to develop a basic quadrotor application for autonomous object tracking in simulation, by integrating computer vision object detection with the MAVSDK MAVLink library for autonomous flight control. PX4 Software in the Loop (SITL) simulation will be used to test the code example. I will talk about how to set up a development environment in the Ubuntu operating system that includes all the required tools-the PX4 SITL simulation environment, the MAVSDK-Python library and the Python OpenCV library. I will then explain a workflow to read and process images to detect objects by using a very basic computer vision technique-color range segmentation. Results obtained from the detection process will then be used to control a simulated quadcopter with the MAVSDK-Python MAVLink library, to make it autonomously track the detected object.