initial_camera_calibration.py

#https://docs.opencv.org/3.3.0/dc/dbb/tutorial_py_calibration.html

import numpy as np
import cv2
import glob
import time

workingdir="/home/pi/Desktop/Captures/"
savedir="camera_data/"

# termination criteria

criteria = (cv2.TERM_CRITERIA_EPS + cv2.TERM_CRITERIA_MAX_ITER, 30, 0.001)
# prepare object points, like (0,0,0), (1,0,0), (2,0,0) ....,(6,5,0)
objp = np.zeros((7*7,3), np.float32)

#add 2.5 to account for 2.5 cm per square in grid
objp[:,:2] = np.mgrid[0:7,0:7].T.reshape(-1,2)*2.5

# Arrays to store object points and image points from all the images.
objpoints = [] # 3d point in real world space
imgpoints = [] # 2d points in image plane.
images = glob.glob('calibration_images/*.jpg')

win_name="Verify"
cv2.namedWindow(win_name, cv2.WND_PROP_FULLSCREEN)
cv2.setWindowProperty(win_name,cv2.WND_PROP_FULLSCREEN,cv2.WINDOW_FULLSCREEN)

print("getting images")
for fname in images:
    img = cv2.imread(fname)
    print(fname)
    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)


    # Find the chess board corners
    ret, corners = cv2.findChessboardCorners(gray, (7,7), None)
    # If found, add object points, image points (after refining them)
    if ret == True:
        objpoints.append(objp)
        corners2=cv2.cornerSubPix(gray,corners, (11,11), (-1,-1), criteria)
        imgpoints.append(corners)
        # Draw and display the corners
        cv2.drawChessboardCorners(img, (7,7), corners2, ret)
        cv2.imshow(win_name, img)
        cv2.waitKey(500)

    img1=img
    
cv2.destroyAllWindows()

print(">==> Starting calibration")
ret, cam_mtx, dist, rvecs, tvecs = cv2.calibrateCamera(objpoints, imgpoints, gray.shape[::-1], None, None)

#print(ret)
print("Camera Matrix")
print(cam_mtx)
np.save(savedir+'cam_mtx.npy', cam_mtx)

print("Distortion Coeff")
print(dist)
np.save(savedir+'dist.npy', dist)

print("r vecs")
print(rvecs[2])

print("t Vecs")
print(tvecs[2])



print(">==> Calibration ended")


h,  w = img1.shape[:2]
print("Image Width, Height")
print(w, h)
#if using Alpha 0, so we discard the black pixels from the distortion.  this helps make the entire region of interest is the full dimensions of the image (after undistort)
#if using Alpha 1, we retain the black pixels, and obtain the region of interest as the valid pixels for the matrix.
#i will use Apha 1, so that I don't have to run undistort.. and can just calculate my real world x,y
newcam_mtx, roi=cv2.getOptimalNewCameraMatrix(cam_mtx, dist, (w,h), 1, (w,h))

print("Region of Interest")
print(roi)
np.save(savedir+'roi.npy', roi)

print("New Camera Matrix")
#print(newcam_mtx)
np.save(savedir+'newcam_mtx.npy', newcam_mtx)
print(np.load(savedir+'newcam_mtx.npy'))

inverse = np.linalg.inv(newcam_mtx)
print("Inverse New Camera Matrix")
print(inverse)


# undistort
undst = cv2.undistort(img1, cam_mtx, dist, None, newcam_mtx)

# crop the image
#x, y, w, h = roi
#dst = dst[y:y+h, x:x+w]
#cv2.circle(dst,(308,160),5,(0,255,0),2)
cv2.imshow('img1', img1)
cv2.waitKey(5000)      
cv2.destroyAllWindows()
cv2.imshow('img1', undst)
cv2.waitKey(5000)      
cv2.destroyAllWindows()