NYCU-Computer-Vision-for-UAV-Autopilot

environment setup

• python=3.9

# install all the required python packages
pip install -r requirements.txt

• pytorch link

Labs

• Lab1

  • Environment Setup (conda virtual environment)
  • OpenCV Basics
  • Image Processing
  • Image Filtering
  • Interpolation
  • Edge Detection

• Lab2

  • Color picture histogram equalization
  • Optimized Otsu Threshold algorithm
  • Connected Component by Two-Pass Algorithm with disjoint set data structure

• Lab3

  • Camera calibartion for camera intrinsic parameters
    • Implement the Zhang's method for camera calibration
  • Wrap perspective transformation for real time camera capture
    • Camera will be mapped to the embedded space of the image
    • Utilizing homography matrix.

• Lab4

  • Marker detection
    • Show the detected markers on the image
    • Show distance between markers and the camera

• Lab5

  • Utilize PID to control the drone
    • Implement the PID controller and keyboard controller for the drone
    • PID controller will help adjust speed to reach the marker

• Lab6

  • Implement the drone autopilot
    • Go to the marker
    • Go right
    • Go to the second marker
    • Go left
    • Land

• Midterm

  • Implement the drone autopilot
    • Marker tracing
    • Marker following
    • Auto rotation, forward, left, right, up, down
    • Auto Land

• Lab7

  • Object detection
    • Using HOG to detect pedestrian
    • Using Haar cascade to detect face
    • Solve PnP with 4 points to get the distance between camera and object