Hi, I’m setting up pose estimation using the aruco module and I’m finding that when I allign the aruco marker centre with the centre of the camera viewport, the tvec X and Y values are not at [0, 0].
Wondering if anyone here has seen similar offsets or has a hunch for why it might be happening.
tvec X at 0: (the purple circle is at [0, 0] on the video frame)
tvec Y at 0:
Apologies for the large code chunks, I’m sharing my camera calibration and pose estimation code in case it’s helpful to see.
# camera calibration
def calibrate():
criteria = (cv2.TERM_CRITERIA_MAX_ITER | cv2.TERM_CRITERIA_EPS, 30, 0.001) # termination criteria
chessboardSquareDimensions = 0.024 # metres
chessboardDimensions = (6, 9)
patternPoints = np.zeros((chessboardDimensions[0] * chessboardDimensions[1], 3), np.float32)
patternPoints[:,:2] = np.mgrid[0:chessboardDimensions[0], 0:chessboardDimensions[1]].T.reshape(-1, 2)
patternPoints = patternPoints * chessboardSquareDimensions
objpoints = [] # 3d point in real world space
imgpoints = [] # 2d points in image place
images = glob.glob('calibration_images_laptop_cam/*.png')
for image in images:
print("next image ")
img = cv2.imread(image)
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
size = gray.shape[::-1]
ret, corners = cv2.findChessboardCorners(gray, chessboardDimensions, None)
if ret == True:
objpoints.append(patternPoints)
corners2 = cv2.cornerSubPix(gray, corners, (11, 11), (-1, -1), criteria)
imgpoints.append(corners)
cv2.drawChessboardCorners(img, chessboardDimensions, corners2, ret)
cv2.imshow('img', img)
key = cv2.waitKey(0)
if key == ord("q"):
cv2.destroyAllWindows()
raise SystemExit
print("")
cv2.destroyAllWindows()
return cv2.calibrateCamera(objpoints, imgpoints, size, None, None)
# pose estimation
arucoSquareDimensions = 0.1 # marker is 10 cm
arucoDict =cv2.aruco.Dictionary_get(cv2.aruco.DICT_4X4_100)
arucoParams = cv2.aruco.DetectorParameters_create()
ret, cameraMatrix, distortionCoefficients, camrvecs, camtvecs = calibrateCamera.calibrate()
print("Sarting video stream..")
capture = cv2.VideoCapture(0)
time.sleep(2.0)
while True:
ret, frame = capture.read()
(corners, ids, rejected) = cv2.aruco.detectMarkers(frame, arucoDict, parameters=arucoParams)
if len(corners) > 0:
ids = ids.flatten()
rvecs, tvecs, markerPoints = cv2.aruco.estimatePoseSingleMarkers(corners, 0.1, cameraMatrix, distortionCoefficients)
