Skip to content

yindaz/PanoBasic

Repository files navigation

Toolbox for panorama image processing

This toolbox provides useful functions to process panorama images. Please see Quick start below for details.

If using this code as parts of your project, please cite:

Y. Zhang, S. Song, P. Tan, and J. Xiao
PanoContext: A Whole-room 3D Context Model for Panoramic Scene Understanding
Proceedings of the 13th European Conference on Computer Vision (ECCV2014)

@article{zhang2014panocontext,
  title={Panocontext: A whole-room 3d context model for panoramic scene understanding},
  author={Zhang, Yinda and Song, Shuran and Tan, Ping and Xiao, Jianxiong},
  booktitle={European Conference on Computer Vision},
  pages={668--686},
  year={2014},
  organization={Springer}
}

If you have any question about the code, please feel free to contact: Yinda Zhang, yindaz at gmail.

Quick start:

Run demo_full.m to see an "almost" complete list of functions in this toolbox. Some (not all) featured functions are:

  • Combine perspective images to a panorama.
  • Project a region on panorama to a perspective image.
  • Line segment detection.
  • Color segmentation.
  • Find vanishing point.
  • Reconstruct 3D cuboid.

For a full list of functions, please see instruction.pdf.

Matterport3D support:

Run demo_matterport.m to see how to stitch panorama for Matterport3D dataset. Note that:

  • If you wish to have seamless good looking color panorama, you should stitch skybox images.
  • If you wish to have aligned color and depth panorama, you should stitch undistorted_color/depth_images.

If you stitch from undistorted_color/depth_images:

  • The color panorama from undistorted_color_images may contain visible artifacts near stitching boundaries because of different exposure between images.
  • The demo only takes the 6 horizontal views as example, and assumes 60 degrees between each pair of adjacent views. This may not be true in practice, and a perfectly accurate panorama stitching requires accurate calculation of (vx,vy) from camera extrinsic matrices.
  • Given more accurate (vx/vy) calculated for upward/downward looking views, you may stitch 18 views together for a more complete panorama.

FAQ:

Q: Windows users will get an error when calling function "lsdWrap.m".
A: The Windows command line does not support '/'. Please change all '/' in commands line to '\' before sending them to "system" function in Matlab.

License:

This toolbox is under the MIT License: http://opensource.org/licenses/MIT.

Releases

No releases published

Packages

No packages published