usage

run main.py file after installing dep pip install -r requirements.txt python3 main.py

Goals

• capture and calibrate camera
• measure distances and direction using ultra sonic and servo
• transform from real world coordinates to camera coordinates
• project the estimated real-world position of an object onto camera image

setup:

• for now lets assume the hardware part exists and works as expected, that means we have:

  • ultrasonic mounted to a servo motor with the needed microcontroller and code
  • get a 1d image from the altrasonic showing the depth of each point

• assuming we have a camera, lol:

  • calibrate usign checkerboard
  • get K, [R|t]

• now we have the equation to transform real world coordinates to image coordinates u=K[R∣t]X

• fusing both real world coordinates with the camera 2d image

• display camera feed, overlay object position, and compare with the visually detected position

---

IO

Input	Output
Servo angle value	Camera feed with projected points shown
Ultrasonic value
Camera feed

NOTES:

• criteria = (type, maxCount, epsilon)
• type = 1 or 2 or 3, using maxcount only, using min acc only, using both
• max iterations, its 30 minimum accuracy 0.001 stop refining if either 30 terations or 0.001 acc achieved
• mgrid shows all available combinations from a 2d points in a n*m grid
  • the grid will return an x, y arrays first shows all rows, second shows all columns and both of size n*m
• cvt frane will be converted from default brg to gray scale image
• findChessboardCorners: as it says and returns the corners
• cornerSubPix will take the corners and refine the corners for higher accuracy based on criteria
• calibrateCamera method to return the calibration matrecis used in transformation 3d to 2d
• drawChessboardCorners: to take the corners and draw them on the camera feed