Skip to content

Latest commit

 

History

History
27 lines (18 loc) · 1.42 KB

Task2.md

File metadata and controls

27 lines (18 loc) · 1.42 KB
Raw

Task 2

Camera calibration

The camera used the task 2 is Lenovo 300 FHD Webcam with FHD 1080P 2.1 Megapixel CMOS Camera and an Ultra-Wide 95 Lens.

  • To calibrate the camera, first print a checkerboard and paste it on a piece of cardboard/wood.
  • Then, run pics.py to click a few images of the board in different angles(around 30). Press 'c' to click the photo and 'q' to end the program when done.
  • Now, update the rows and columns in camera.py according to the size of the board.
  • Do not include the outermost lines when counting the rows and columns.
  • Now, make appropiate changes in path of the image folder(if any) in camera.py and run the code.
  • It will return a JSON file containing the camera matrix and the distortion coefficients.
  • Use those matrices in marker.py in Camera directory.

Pose Estimation

The pose of the drone is obtained from the Aruco marker attached on the drone. The cv2.aruco library, available in the opencv-contrib-python package, is used to detect the drone and its pose, specifically it's x-y coordinates in pixels, height in meters and yaw in radians.

PID control

Task 2

Running

To run the PID control task, run master.py which will further start two parallel processes for pose estimation and PID control.