Brummi / MonoRec

Official implementation of the paper: MonoRec: Semi-Supervised Dense Reconstruction in Dynamic Environments from a Single Moving Camera (CVPR 2021)

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

the setup of KITTI dataset

ZheningHuang opened this issue · comments

Hi, I was trying to reproduce the point clouds for KITTI Odometry dataset with the pre-trained model you provided. I am a bit confused about the setup of KITTI dataset for this.

You mentioned

"To setup KITTI Odometry, download the color images and calibration files from the official website (around 145 GB).

I wonder if this 145 GB is comprised of:
Download odometry data set (color, 65 GB)
Download odometry data set (velodyne laser data, 80 GB)?
at the official website.

For the task I described, (not to train the model, just to reproduce the point clouds), I have done the following:

  1. download odometry data set, including color (65 GB) and calibration files (1 MB) from the
  2. Replace the calib.txt and times.txt files in color (65 GB) with calib.txt and times.txt files in calibration files (1 MB)
  3. Download the pre-trained final model by running download_model.sh
    From the error I got, it seems I still missing '../../../data/dataset/sequences/00/image_depth_annotated'

Would be really appreciate it if you could let me know how can I got this. If possible, also correct me if I did anything wrong.

Thanks for your help

Hi! How did you solve the problem? I just want to rebuild the cloud (no training required) too. Do I have to download the entire dataset? Can you share your steps~ Thanks a lot!!