{
  "id": 405407,
  "title": "Set 1 camera's rotation and translation to 0?",
  "url": "/competitions/image-matching-challenge-2023/discussion/405407",
  "author_name": "",
  "post_date": "2023-04-27T12:29:06.996098300Z",
  "votes": 6,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Hi all,<br>\nAs the rotation and translation is measured relative between each possible pair of cameras (and not a universal frame of reference) then does that mean it makes the most sense to set 1 camera's location/translation arbitrarily, to use it for the \"anchor\" to base the rest of the camera positions on? For instance, setting all values in the rotation matrix and translation vector to 0? I am struggling to see why this would not be the correct approach, and how else one is supposed to get the values for each camera if all camera position information is relative.<br>\nThanks</p>",
  "messages": [
    {
      "id": "2237171",
      "postDate": "04/27/2023 12:29:06",
      "content": "<p>Hi all,<br>\nAs the rotation and translation is measured relative between each possible pair of cameras (and not a universal frame of reference) then does that mean it makes the most sense to set 1 camera's location/translation arbitrarily, to use it for the \"anchor\" to base the rest of the camera positions on? For instance, setting all values in the rotation matrix and translation vector to 0? I am struggling to see why this would not be the correct approach, and how else one is supposed to get the values for each camera if all camera position information is relative.<br>\nThanks</p>",
      "rawMarkdown": "Hi all,\nAs the rotation and translation is measured relative between each possible pair of cameras (and not a universal frame of reference) then does that mean it makes the most sense to set 1 camera's location/translation arbitrarily, to use it for the \"anchor\" to base the rest of the camera positions on? For instance, setting all values in the rotation matrix and translation vector to 0? I am struggling to see why this would not be the correct approach, and how else one is supposed to get the values for each camera if all camera position information is relative.\nThanks",
      "votes": null
    },
    {
      "id": "2244044",
      "postDate": "05/03/2023 11:49:32",
      "content": "<p>Yes, it does not matter. You can set it to zeros or choose arbitrary values.</p>",
      "rawMarkdown": "Yes, it does not matter. You can set it to zeros or choose arbitrary values.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2244044,
      "author_name": "eduardtrulls",
      "author_url": "",
      "post_date": "05/03/2023 11:49:32",
      "content": "<p>Yes, it does not matter. You can set it to zeros or choose arbitrary values.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2237171": "Hi all,\nAs the rotation and translation is measured relative between each possible pair of cameras (and not a universal frame of reference) then does that mean it makes the most sense to set 1 camera's location/translation arbitrarily, to use it for the \"anchor\" to base the rest of the camera positions on? For instance, setting all values in the rotation matrix and translation vector to 0? I am struggling to see why this would not be the correct approach, and how else one is supposed to get the values for each camera if all camera position information is relative.\nThanks",
    "2244044": "Yes, it does not matter. You can set it to zeros or choose arbitrary values."
  },
  "source": "meta"
}