<br> [MAS.865](../..) &gt; [Scanning Processes](..) &gt; Photogrammetry ## Photogrammetry <div style="margin-left:2.5%"> [Photogrammetry](https://en.wikipedia.org/wiki/Photogrammetry) is the science of making measurements from photographs. ### Basics Types of 3D reconstructions, <strong>Shape from X</strong> where X is * Shape from <strong>Motion</strong>, also known as <strong>Structured from motion</strong> (SfM) * Shape from <strong>Shading</strong>, * Shape from <strong>Texture</strong> * Shape from <strong>Focus/Defocus</strong> * Shape from <strong>Specularity</strong> * Shape from <strong>Shadows</strong> * ... See these overviews: * [CVG @ ETHZ](http://www.cvg.ethz.ch/teaching/2012spring/3dphoto/Slides/3dphoto12_shapeFromX.pdf) * 6.869 @ MIT * [Image formation](http://6.869.csail.mit.edu/fa13/lectures/lecture10imageformation.pdf) * [Shape from X](http://6.869.csail.mit.edu/fa13/lectures/lecture11shapefromX.pdf) ### Structure from Motion Stereo (two views): ![Stereo](shape_from_stereo.png) ![Epipolar](epipolar_geometry.png) * Epipolar geometry * Feature matching Multiview (N views): ![Motion](shape_from_motion.jpg) * [Phototourism](http://phototour.cs.washington.edu/) * [Video](https://www.youtube.com/watch?v=IgBQCoEfiMs) * Bundle Adjustment (simultaneous geometry and camera refinement in SfM) * [Bill Triggs, Philip McLauchlan, Richard Hartley and Andrew Fitzgibbon](http://lear.inrialpes.fr/pubs/2000/TMHF00/Triggs-va99.pdf) * [Bundler: SfM for Unordered Image Collections](http://www.cs.cornell.edu/~snavely/bundler/) * [BigSFM](http://www.cs.cornell.edu/projects/bigsfm/) * Building Rome in a Day * Hyperlapse * [First-Person Hyperlapse](https://blogs.microsoft.com/ai/hyperlapse-siggraph-2014/) * [Visual Structured from Motion](http://ccwu.me/vsfm/) * [Towards Internet-scale Multi-view Stereo](https://www.microsoft.com/en-us/research/publication/towards-internet-scale-multi-view-stereo/), CVPR 2010 * [Pix4D](https://pix4d.com/) ### Shape from Shading ![Shading](shape_from_shading.gif) ![Normals](deep_normals.png) * [Shape from Shading: A Method for Obtaining the Shape of a Smooth Opaque Object From One View](http://people.csail.mit.edu/bkph/AIM/AITR-232-OPT.pdf), Horn, 1970 * Assumes Lambertian surface apperance model * Integration of normals from shading * [Slides](http://thaines.com/content/research/sfs_pres.pdf) from Tom Haines, 2012 * [Assumed Light Direction for Perceiving Shape from Shading](http://vis.berkeley.edu/papers/assumedlighting/), O'Shea, Banks and Agrawala, APGV 2008 * Recently done with Deep CNNs * [Designing Deep Networks for Surface Normal Estimation](http://www.cs.cmu.edu/~xiaolonw/papers/deep3d.pdf) #### Intrinsic images: decomposing shading and reflectance. ![Intrinsics](intrinsic_images1.png) ![Instrinsics](intrinsic_images2.png) * [Papers / UCBerkeley](https://www2.eecs.berkeley.edu/Research/Projects/CS/vision/reconstruction/) * [Intrinsic Images in the Wild](http://opensurfaces.cs.cornell.edu/intrinsic/) * [Ground truth dataset and baseline evaluations for intrinsic image algorithms](http://www.cs.toronto.edu/~rgrosse/intrinsic/), Grosse, Johnson, Adelson and Freeman, ICCV 2009 #### From Specularities: ![Specularity](shape_from_specularity.png) ![Glossy](shape_from_glossy.jpg) * [Depth Estimation for Glossy Surfaces with LF Cameras](https://cseweb.ucsd.edu/~ravir/l4cv.pdf), Tao et al., ECCV 2014 * [Shape from Specular Flow](https://ieeexplore.ieee.org/document/5499479/), Adato et al., TPAMI 2010 * [Secular reflections and the perception of shape](http://persci.mit.edu/pub_pdfs/Fleming_jov-4-9-10.pdf), Fleming, Torralba and Adelson, Journal of Vision 2004 ### Shape from Texture ![Texture](shape_from_texture.png) * [Examples](http://6.869.csail.mit.edu/fa13/lectures/lecture11shapefromX.pdf#page=89) * [Shape from Texture, J. Aloimonos 1988](https://link.springer.com/article/10.1007/BF00363944) * Texton, as the "putative units of pre-attentive human texture perception" * [Textons, the elements of texture preception, and their interaction](https://www.nature.com/articles/290091a0), Bela Julesz, Nature 1981 ### Shape from Focus / Defocus ![Defocus](shape_from_defocus.png) * [Shape from Focus System](http://graphics.stanford.edu/courses/cs348b-06/homework3/Nayar_CVPR92.pdf), Shree K. Nayar, CVPR 1992 * [Shape from Focus](https://dl.acm.org/citation.cfm?id=628640), Nayar and Nakagawa, TPAMI 1994 * [Depth from Shading, Defocus, and Correspondence Using Light-Field Angular Coherence](https://cseweb.ucsd.edu/~ravir/cvpr_shading.pdf), Tao et al., CVPR 2015 * [Depth from Semi-calibrated stereo and defocus](https://cseweb.ucsd.edu/~ravir/semistereo.pdf), Wang et al., CVPR 2016 ### Shape from Shadows ![Shadow](shape_from_shadows.png) * [Shape From Shadows](https://scholar.harvard.edu/pcavanagh/publications/shape-shadows), Cavanagh and Leclerc, Journal of Experimental Psychology, 1989 * [On 3D surface reconstruction using shape from shadows](https://ieeexplore.ieee.org/document/698646/), Daum and Dudek, CVPR 1998 ### Miscellaneous * https://www.myminifactory.com/category/scan-the-world * https://www.artmyn.com/