Skip to content
This repository was archived by the owner on Apr 4, 2025. It is now read-only.

Latest commit

 

History

History
133 lines (78 loc) · 5.42 KB

RealityCaptureToColmap.v2.md

File metadata and controls

133 lines (78 loc) · 5.42 KB

RealityCapture to Colmap to Postshot Script

This script was made to simplify the process from this guide: https://gist.github.com/jo-chemla/258e6e40d3d6c2220b29518ff3c17c40 by jo-chemla

This repository contains a Python script designed to automate the process of converting images and data from RealityCapture to a format compatible with Postshot via Colmap.

Prerequisites

  • Ensure you have Python installed on your system, and the pip package manager is functional.
  • Install Colmap on your system. You can download it from the Colmap github.
  • Ensure that colmap.exe and colmap.bat are exposed on your computer's PATH environment variable. This means that you should be able to run colmap from any directory in the command line.

Example paths for colmap exe and bat (depending on your own installation path you need to define):

(location of colmap.bat)

  • E:\Program Files\Colmap\COLMAP-3.9.1-windows-cuda

(location of colmap.exe)

  • E:\Program Files\Colmap\COLMAP-3.9.1-windows-cuda\bin

These two paths should be added to your system's PATH as environment variables.

Installation

Clone this repository to your local machine using:

git clone https://github.com/Maxiviper117/RealityCapture-to-Postshot

Navigate to the cloned directory:

cd RealityCapture-to-Postshot

Exporting Files from RealityCapture

Export Step 1: the Bundler .out File

Choose a folder to store the exported file.

  1. Set Fit option to Inner Region.
  2. Set Resolution option to Fit.
  3. Ensure Export Images is set to No.
    • We will export the images separately in the next step when exporting the Undistorted images with image list. export option

alt text

Export Step 2: the Undistorted Images with Image List

Choose a folder to store the exported file.

  1. Set Fit option to Inner Region.
  2. Set Resolution option to Fit.
  3. Under Export image settings, choose the desired image format (preferably jpg or jpeg).
  4. Set Naming Convention to original file name.
  5. Enable Customize image path and select the folder to store images.
  6. (Optional) Adjust the Downscale option if required.

alt text

Running the Script

Configuration

The script requires the following directory arguments:

  • --images_dir: The path to the folder containing the undistorted images. (The folder selected in the previous step for the undistorted images)
  • --working_dir: The path to the folder containing the .out and .lst files. (The folder selected in the first step)
  • --output_dir: (OPTIONAL) - The path to the folder where the script will output the Colmap workspace. (defaults to a folder named colmap-workspace in the working directory)

Example Command

NOTE: Must run the script from the command line as an administrator.

py RealityCaptureToColmap.v2.py --working_dir "<full path to working directory from previous export step 1>" --images_dir "<full path to images folder from previous export step 2>"

Option choose the output directory:

py RealityCaptureToColmap.v2.py --working_dir "<full path to working directory from previous export step 1>" --images_dir "<full path to images folder from previous export step 2>" --output_dir "<full path to output directory>"

Execution

Run the script from the command line. It will install the necessary kapture package into your Python environment and process the provided directories.

Outputs

Upon successful execution, the script outputs a colmap-workspace (or the specified output directory) folder in the working directory with subfolders for different dataset stages. The dataset-colmap folder contains the necessary files for Postshot import. Look for the .bin files in dataset-colmap\sparse\0.

alt text

Importing to Postshot (Option 1)

Copy the all three cameras.bin, images.bin and points3D.bin file (from the <output_dir>\dataset-colmap\sparse\0 folder) and the images (exported from RealityCapture) into the same folder. (Either all three bin files in that same images folder or all images and bin files in a new folder)

Then simply drag and drop the folder into Postshot to start the import process.

You should see this pop-up window, showing Camera Poses set to import. (this means the that it imported the cameras.bin file successfully), you can now start training.

alt text

Failed to import camera poses error.

If you see this error, when you drag and drop the folder into Postshot, try close and open Postshot again, and drag and drop the folder again.

Importing to Postshot (Option 2)

Another way to import the data into Postshot is to first import the images separately, when the pop-up window appears, uncheck option start training then click import.

Then select the image set, then at the bottom right corner, under Actions fold, click on Import Camera Poses and select the cameras.bin file.