Loading...

Using Visual Patterns for Precise Automatic Landing of a Quadcopter

Noorian, Hamid Reza | 2020

423 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 53883 (45)
  4. University: Sharif University of Technology
  5. Department: Aerospace Engineering
  6. Advisor(s): Nobahari, Hadi
  7. Abstract:
  8. In this study, the method of recognizing image patterns by image processing to guide the quadcopter to the landing point is expressed. For this, first the forces and torques on the quadcopter in the body system are calculated and after transferring to the inertial coordinates, the quadcopter dynamics with the use of Newton-Euler law is simulated in the MATLAB software environment. Then, a simple PID controller is designed to control the attitude, height and position of the quadcopter and create stability in the quadcopter dynamics. Because quadcopter navigation is dependent on visual data, a simulation environment called VRML is used. This environment has the ability to create a three-dimensional space for the quadcopter to move, as well as creating a street-like environment with the desired image patterns. The OpenCv library in Python was used to detect these visual patterns. One of the main challenges in this project was to create a connection between two different software environments such as Python and MATLAB, which was done completely. The two main goals of this project are to control the yaw angle of the quadcopter using a simple visual pattern such as arrows, as well as the use of nested circles for accurate landing until the last moment, which is well done and is one of the innovations of this project. The image processing algorithms used in this project are color recognition and edge recognition algorithms
  9. Keywords:
  10. Quadcopter ; Color Detection ; Image Processing ; Image Edge Detection ; Automatic Landing ; Percise Landing

 Digital Object List

 Bookmark

No TOC