isrishtisingh / tello-drone-programming Goto Github PK

• The aim of this project was to program a tello drone to perform basic computer vision tasks such as object detection, face detection, etc.
• Programming was done using the OpenCV library and Python API called djitellopy which provides an easy interface over the official Tello SDK and Tello EDU SDK.
• The project contains many smaller modules and computer vision tasks implemented to control the DJI Tello.
• It is an ongoing project of mine, more modules coming soon integerating more computer vision and deep learning in the drone.
• References and inspiration were taken from the lecture series by Murtaza's Workshop - Robotics and AI

In this module, the aim was to program the drone to:

• Take live input from its surrounding
• Extract out frames from it.
• Detect face using opencv’s state of the art face detector and return a score for the face depending on the size of it visible in the scene and its orientation
• Based on the above scores, decide weather to move away or towards the face and decide the rotation movement based on orientation of the face. -Performed reasonably well and was able to track a given subject.

In this module, the aim was to program the drone to:

• Given a particular path, trace the path, detect the path and follow along the path.
• The path can be really complex consisting of complicated curves and angles.
• Motion is with a controlled angular and linear velocity which is also programmed.
• Motion is delayed to have grater control and smoother mapping.

In this module, the aim was to program the drone to:

• Move around in the scene, keep capturing live input obtained from the scene and store the feed.
• The drone can perform movement in 12DOF at a particular speed which can be controlled.

• Basic Movements: The drone can be controlled to perform various basic movements at different directions and DOFs.
• Image capture : The drone can capture images at various orientations, extract and store the inputs for further processing.
• Keyboard control: The drone can be controlled using the system keyboard instead of an RC(incase RC is not available) to control its movement, flight, landing etc.

• https://github.com/damiafuentes/DJITelloPy
• https://youtu.be/LmEcyQnfpDA