Improve robustness of UR hand-eye sample with regards to missdetection

[SatjaSivcev]

If the calibration object cannot be detected in a single zdf from the whole dataset, the code sample fails. Add code to enable skipping a pair of transform and point cloud in case detection fails (we have this in the CLI tool).

https://github.com/zivid/zivid-python-samples/blob/508f22b06bc6d46364de66e4b6463fd515cc9211/source/applications/advanced/hand_eye_calibration/ur_hand_eye_calibration/universal_robots_perform_hand_eye_calibration.py#L400-L401