StereoCameraReconSim

Framework for Simulation of 3D mesh reconstruction with given stereo pinhole camera setup.

Overview

StereoCameraReconSim is a simulation framework in MATLAB for a common main objective surgical mircoscope using multiple cameras. Deepending on a stl input mesh, it provides for each camera the respective image, the complete 3D reconstructable mesh, depending on a visibility threshold and reconstruction area in quantitative number. For a better interpratation it is possible to excluse vertices from the reconsturction. The framework is based only on the "Parallel Compunting Toolbox" for acceleration. For deeper information, see the paper.

Configuration

There are different files for the simulation configuration:

1. simSettings.m Config parameter:

   • specify program parameters

     • headless: Set true if you intend to run this program without a gui
     • suppressWarnings: Set false to unsuppress all warnings

   • input mesh

     • filename: Specify stl file name (file must be in the same folder as "simSettings.m" or in one of the subfolders)
     • indexZero: So the vertex indices in your stl start with 0? If so: set true. Use false if your vertex ID list was taken from MATLAB.

   • reconstruction parameter

     • reconThreshold: Specify minimum number of cameras that must see a surface patch for it to be considered reconstructable

   • saving & output parameters

     • savePlots: Set true to save matlab figures. NOTE: use figure(openfig("filepath with extension")) to open the saved plots
     • saveWorkspaceVariables: Set true to save the whole workspace. NOTE: the output file could be several/many GB large and saving might take considerable time.

   • plot settings

     • displayPlots: Set true to display the plots.
     • enforcePlot: Set true to force plotting all camera images even if there are more than 6 cams (NOTE: this can cause laggy behavior and require very long time to open all the figures)
     • showIDs: Set true to show face IDs in plots. this is intended for debugging purposes only and will clutter your plots greatly. Plots will become very laggy.
     • showNormals: Set true to show normals in plots

2. camSetup.m Config parameter:

   • camera setup parameters

     • camNum: Specify the number of cameras
     • center: Specify the point which will be the hypothetical "focal point" of your camera setup's objective lens.

   • parameters of the ring on which the cameras lie:

     • L: Hypothetical working distance L (between mesh and camera ring)
     • b: Baseline (ring diameter for positioning the cameras)

   • intrinsic camera parameters

     • L_x,L_y: Sensor dimensions (width, height) in 'mm'
     • N_x,N_y: Sensor dimensions (width, height) in 'pixels'
     • o_x,o_y: Principal point offset in 'pixels'

3. exclude_ids

   • Contains a mat-file with the vertec ids wich were excluded from the calulation. (NOTE: detailed description in the identically named folder.)

Run

After setting the configuration parameters:

Just run "runSim.m" in MATLAB.

Database configuration

StereoCameraReconSim
│   camSetup.m 			  # Config file for the camera setup
│   runSim.m 			  # Script for starting the simulation
│   simSettings.m 		  # Config file for programm settings
│
├───exclude_ids
│       readme_exclude_faces.txt  # Discription for futher information
│
├───functions		          # contain the core functionallity
│   │   ...
│   │
│   └───helpers
│       │   ...
│       │ 
│       └───third_party
│               ...
│
└───stl
    └───examples		  # Example meshes
	    50x50_cube.stl
	    center_cone.stl
	    nested_cubes.stl
	    thing.stl

License

All source code is subject to the terms of the GNU General Public License v3.0.
Copyright 2020 Jan Kost, Andreas Wachter - Karlsruhe Institute of Technology.

Citation

If you use this software, please cite it as follows:

Wachter A, Kost J, Nahm W. Simulation-Based Estimation of the Number of Cameras Required for 3D Reconstruction in a Narrow-Baseline Multi-Camera Setup. Journal of Imaging. 2021; 7(5):87. https://doi.org/10.3390/jimaging7050087

Contact

Jan Kost, Andreas Wachter Institute of Biomedical Engineering
Karlsruhe Institute of Technology
www.ibt.kit.edu