CALIB_HAND_EYE: _ANDREFF & _DANIILIDIS fail #17986
Description
System information (version)
- OpenCV => 4.4
Detailed description
using DS1.zip from:
https://data.nal.usda.gov/dataset/data-solving-robot-world-hand-eyes-calibration-problem-iterative-methods
- CALIB_HAND_EYE_ANDREFF returns a NaN rotation
- CALIB_HAND_EYE_DANIILIDIS returns a implausible transform
Steps to reproduce
import cv2
import numpy as np
corners = []
cap = cv2.VideoCapture("DS1/images/camera0/image%d.png")
while True:
ret, img = cap.read()
if not ret: break
corners.append(cv2.findChessboardCorners(img, (8, 6))[1])
objp = np.zeros((6*8, 3), np.float32)
objp[:, :2] = np.mgrid[0:8, 0:6].T.reshape(-1, 2) * 28.5 # [mm]
objpts = [objp]*len(corners)
rvecs, tvecs = cv2.calibrateCamera(np.array(objpts), corners, (640, 480), None, None)[3:]
rposes = np.loadtxt("DS1/robot_cali.txt", skiprows=1).reshape(-1, 4, 4)
for i in range(5):
print(cv2.calibrateHandEye(rposes[:, :3, :3], rposes[:, :3, 3], rvecs, tvecs, method=i)) Issue submission checklist
- I report the issue, it's not a question
- I checked the problem with documentation, FAQ, open issues,
answers.opencv.org, Stack Overflow, etc and have not found solution - I updated to latest OpenCV version and the issue is still there
- There is reproducer code and related data files: videos, images, onnx, etc