Meshroom: Initial Pipeline, CCTags, using a Turntable and Known Camera Positions

Поделиться
HTML-код
  • Опубликовано: 17 ноя 2024

Комментарии • 10

  • @sanyammishra869
    @sanyammishra869 5 дней назад +1

    Best video I saw on mesh room. Thank you for making this video 😁❤... Can you please give me some more information on exactly how you made your own nodes or provide some reference material to figure it out.

  • @ukaszpruszynski1799
    @ukaszpruszynski1799 18 дней назад

    Thank you! Best meshroom video I've seen. Epic work!

  • @sergiovr3
    @sergiovr3 9 дней назад

    You are the best!. This video have more quality than a lot of paid courses, you gave me the motivation to keep studing meshroom. If you have a donation link I would love to contribute
    PS: I use a closed white cardboard box with 2 lights and it seems that free me from a lot of problems with the background that you describe in the video

  • @eschersky
    @eschersky Месяц назад +3

    Great information, thank you so much for making this! Most tutorials on Meshroom I found are not very detailed, so apart reading the manual your video is the best thing to finally start understanding the advanced settings!

    • @_mpr_
      @_mpr_  Месяц назад

      Thank you! That's what I was going for :)

  • @eastonpotokar6725
    @eastonpotokar6725 10 дней назад

    This is a great tutorial, thanks so much for sharing. So much info on meshroom here that's undocumented elsewhere.
    Would you mind sharing your custom nodes? Just to jumpstart someone who's trying to do something similar

  • @RebelPhoton
    @RebelPhoton Месяц назад +2

    Awesome!

  • @fuff8147
    @fuff8147 Месяц назад

    Can't get CCTags to work properly, the FeatureExtraction node behaves very oddly for me. When enabling cctag3 or cctag4, the node crashes on compute (nothing in the log, "aliceVision_featureExtraction has stopped working" from Windows) and no features are extracted. This only happens when I try to load more than 4 images. With 4 images it works fine and detects the tags, but as soon as I put in one more image it crashes on compute. Am I missing something obvious?

    • @_mpr_
      @_mpr_  Месяц назад +1

      If it crashes like that then there's probably an issue in the specific binary of Meshroom that you're using. It also took me some time (an eternity) to get it set up so that everything works. I had to build it in a Docker image that uses Debian 12.6 as base image, I couldn't get it to run on reliably on Arch Linux, Ubuntu 22.04 or Ubuntu 24.04. If you were on linux then you could use the Dockerfile that I linked to in the video description. Unfortunately I can't help you with Windows, haven't used that in a long time ...

    • @fuff8147
      @fuff8147 Месяц назад

      @@_mpr_ Linux it is then, time to set up a Debian VM... and I should create an issue on GitHub. In the meantime I'm using MeshroomCL for datasets that need CCTags, runs considerably slower but at least it doesn't crash! Thank you for the fantastic tutorial by the way, great intro to photogrammetry with Meshroom. You answered all the questions I had in one concise video. Grüß Gott aus Norddeutschland!