The bedroom of a game developer living out in the very heart of the English countryside. The building was built in the early 1900s and the loft was used for the servants' quarters. I guess they were either short or used to keeping their heads down - either way the door frames have had a fair amount of skulls smack against them since.

Generated with Agisoft Photoscan, then cleaned up in Meshlab, Maya and Houdini.

White Hart House

Texture mapping on the clandestinely scanned bust
--
The texture was retrieved from my previous [photogrammetry model](https://skfb.ly/GWMR), while the 3D object was downloaded from [AD&amp;D 4D](https://skfb.ly/6QUFB)' Sketchfab model. 

The combination of these two models was completed with **MeshLab**, using the very efficient "Point Based Gluing" alignment method.

Finalization with realistic lightings performed in Sketchfab.

Click here to download the texture (PNG format, 1024 &times; 1024 pixels): http://c.yamahata.fr/sketchfab/Nefertiti_texture_HQ2.png

---
Check [Nefertiti Hack](http://nefertitihack.alloversky.com/) to get the original _.obj_ high resolution model (by Jan Nikolai Nelles & Nora Al-Badri).

🎞️ You may also check out this [Youtube video](https://youtu.be/x7cV316ulGY?si=g57Q1V83t7QR6q-O). I made it by combining footages from the Neues museum in Berlin with this 3D model.


![](https://i.imgur.com/BYjUyfw.png)

*Texture (PNG format, 1024 &times; 1024 pixels)*

Nefertiti's bust (like in the museum)

Learn how to use open source Meshroom in this step by step tutorial on the Sketchfab Blog:
https://sketchfab.com/blogs/community/tutorial-meshroom-for-beginners

Low-poly version of the [default pipeline result](https://skfb.ly/6HNN9). The original model has been simplified using an automatic decimation process within Meshroom, to limit the number of vertices to a maximum of 100K.

It has been cleaned up in Meshlab (bounding box cut-out) and retextured in Meshroom.

**Details**

 - Meshroom 2019.1.0
 - Input: 220 images
 - Camera: Sony A7S II

Buddha - 100K Vertices Decimation

This model is the result of Meshroom default photogrammetry pipeline. 
It has been cleaned up in Meshlab (bounding box cut-out) and retextured in Meshroom.

**Details**

- Meshroom 2019.1.0
- Input: 220 images
 - Camera: Sony A7S II

Buddha - Default Pipeline

This is the biggest model I have posted in Sketchfab, that could make it slow to load. This model of the masterchief monument - executed by French sculptor Aimé-Jules Dalou, is made from almost 500 photos of 24 MPixel (taken from the ground) with ContextCapture, the photogrammetry software of Acute3D, Bentley Systems. I did not retouch the mesh, only cropped it, remove texture, made ambient occlusion in Meshlab, used lighting of Sketchfab.

The 3D model with texture can be viewed here: http://goo.gl/HURMmB

For information about this monument, I refer you to the wiki page (in French) which can be translated in English with Google translation: https://fr.wikipedia.org/wiki/Le_Triomphe_de_la_République. 

A heavily downsampled model of this model, which is downloadable is here : https://skfb.ly/JS8x

Place de la Nation-The Triumphe of the Republic

[Original HR object](https://www.artec3d.com/3d-models/napoleon) (6.8M tris) captured by [Artec 3D](https://www.artec3d.com/) ( [on sketchfab](https://skfb.ly/RWsK) - 200k tris).

![Imgur](https://i.imgur.com/8kJkZgs.jpg)

# Process

* The initial mesh was remeshed to an intermediate object with Meshlab (quadric edge collapse decimation) and **mmgs** (6.8M -> 950k tris)
* Details (ropes and sword) were splitted from the horse body (mark sharp edges -> edge split -> separate), and batch remeshed for better results.
* The resulting meshes (70k tris total) were then manually unwrapped (+manifold edges seams) and merged.

# Material

The material was created from scratch on the high resolution object before being baked to 2K textures:

* Bronze patina on the top
* Mix of metals on the body
* Smoother metal on the edges

 &nbsp;

*Cycles material layout*

![](https://i.imgur.com/BBpRJg7.jpg)

*Blueprint*

![](https://i.imgur.com/p0yLYdo.jpg)

Equestrian statue of Napoleon

Calatrava La Vieja is a Medieval fortress located in Carrión de Calatrava (Castilla-La Mancha,  Spain). The site was occupied as early as the 8th Century A.D. and through the early 15th Century by both Christians and Muslims at different periods. Calatrava La Vieja  was the home to the first military order of the region: the Order of Calatrava. The model was made using 2,442 images captured with a DJI drone; images were processed in Reality Capture software with additional processing in Geomagic and Meshlab.

Calatrava La Vieja (Ciudad Real, Spain)

Scanned with the Heges app (https://hege.sh) on my iPhone X. Post-processed in Meshlab (cut-out the mouflon from background & removed noise).

Be sure to check out mesh quality!

Mouflon | Scanned with Heges app on iPhone X

This model is the result of draft meshing directly from Structure-from-Motion output. It has been cleaned up in Meshlab (bounding box cut-out) and retextured in Meshroom.

**Details**

- Meshroom 2019.1.0
 - Input: 220 images
 - Camera: Sony A7S II

Buddha - Draft Meshing from SfM

This iconic bust of Nefertiti was reconstructed by **photogrammetry**, _solely_ from images found on the web. Hundreds of photographs can easily be found (_e.g._, Google images, flickr) ... but only a handful of photographs are available for the back and the top of the bust. This explains why the reconstruction is not so good for these sides. Fortunately, lighting can be adjusted in Sketchfab and the final result is rather satisfying considering the limited resources used to build this 3D model.

Softwares used: **VisualSFM**, **MeshLab**.
See the YouTube tutorial _["3D scanning for free!"](https://youtu.be/D6eqW6yk50k "3D scanning for free!")_ by [Phil Nolan](https://sketchfab.com/philnolan3d "Phil Nolan") to learn how to use these free softwares.


![](https://i.imgur.com/9isCEZd.png)
  _**Screenshot of VisualSFM**: Application used for this 3D reconstruction from photographs [found on the web](https://www.google.ch/search?q=nefertiti+bust&tbm=isch)._

Nefertiti's bust

Castillo de Alhama de Murcia is located in the Murica region of Spain. The castle is situated high on a steep hilltop and is of Islamic origin, built during the 12th and 13th centuries. Portions of the castle recently have been restored. Important structural features include a main tower, extensive wall fortification, and aljibes (cisterns). This 3D model was made using 1590 images captured with a drone and 25 FAROFocus3D scan positions (tower interior) that were processed in Reality Capture, Geomagic Wrap, and Meshlab.

Castillo de Alhama de Murcia

This macaque skull was scanned using CT. A 3D model was created using 3D Slicer, and cleaned using Meshlab. This project was made possible by a University of Melbourne Learning and Teaching Initiatives Grant.

Macaque Skull

Neoclassical bridge built in the second half of the 18th century by one of the best Spanish architects: Juan de Villanueva (1739-1811). This bridge was part of what is known as Canal de Gran Prior, a large hydraulic infrastructure that sought to achieve better use of the water from the Guadiana river in La Mancha. After decades of abandonment, it was excavated and recovered in 2021. It is currently preserved intact a few kilometers from the small town of Alameda de Cervera (Alcázar de San Juan, Spain).

Processed in Reality Capture from 4058 images. Additional processing completed in Geomagic Wrap, Meshlab, and Blender.

This work is part of the collaboration agreement signed in 2020 between the Junta de Comunidades de Castilla-La Mancha and Global Digital Heritage for the digitization of the Cultural Heritage of Castilla-La Mancha region.

Vado Lancero Neoclassical Bridge (Alameda)

Neoclassical bridge built in the second half of the 18th century by one of the best Spanish architects: Juan de Villanueva (1739-1811). This bridge was part of what is known as Canal de Gran Prior, a large hydraulic infrastructure that sought to achieve better use of the water from the Guadiana river in La Mancha. It is currently preserved intact in the small town of Alameda de Cervera (Alcázar de San Juan, Spain).

Processed in Reality Capture from 3082 images. Additional processing completed in Geomagic Wrap, Meshlab, and Blender.

This work is part of the collaboration agreement signed in 2020 between the Junta de Comunidades de Castilla-La Mancha and Global Digital Heritage for the digitization of the Cultural Heritage of Castilla-La Mancha region.

Neoclassical Bridge (Alameda de Cervera, Spain)

This model shows one of two prison cells of the town of Picón (Ciudad Real, Spain). Its construction dates from the 16th century, and it is one of the few examples that are preserved in the La Mancha region.

This model was created from 17 scans from a FARO Focus X330 laser scanner. These data were aligned and exported in FARO SCENE and the meshing and final processing was in Meshlab.

The data for this project were collected in 2017 as a joint effort between the University of Castilla-La Mancha (UCLM), the Fondazione Bruno Kessler (FBK) and CVAST at the University of South Florida (USF) under the legal permission of Junta de Comunidades de Castilla-La Mancha (Expte. 16.0834-P1. Sistema de documentación 3D del Patrimonio Cultural de Castilla-La Mancha). The project is now managed by UCLM and Global Digital Heritage (GDH). UCLM project co-Director Dr. Víctor Manuel López-Menchero Bendicho

Prison Interior (B) Picón (Ciudad Real, Spain)

My next entry for the #AgisoftClothesChallenge

This jeans jacket was reconstructed in Agisoft Metashape from 648 21Mpx photos all of which aligned perfectly.

I improved the previous setup with better weights and smaller strands [to not obscure the view], I also took the pictures on a cloudy day with no wind. This time I had no problems with alignment.

The model was made from depth maps instead of dense cloud, I came to conclusion that depth maps recreate small details so much better than dense cloud [I assume this is why reality capture doesn't have dense cloud stage], it was  also a bit faster than dense cloud generating. 

The original model had 7,151,887 faces so I decimated it to 100k, removed some flying noise in Meshlab and in the end I delightened the texture in Agisoft Delighter.

Link to non-delightened texture: https://drive.google.com/file/d/1EnPlgy1SaJGJPwoC7c56FM60aWC9BkK_/view?usp=sharing

Jeans Jacket #AgisoftClothesChallenge

A skull point cloud generated from a CT scan of the [Visible Human Female](https://www.nlm.nih.gov/research/visible/animations.html) downloaded from [here](https://mri.radiology.uiowa.edu/visible_human_datasets.html).

CT .nrrd volume generated with 3D Slicer was imported into Paraview. A color ramp was applied and the volume was "clipped" to the voxels representing mostly bone then converted to a point cloud. Meshlab was used to clean up extraneous points. The volume of point clouds looks similar to the volume rendering in CT software but can be rendered in Sketchfab.

YouTube video explaining the process: https://youtu.be/00JkLcklOpc.

VHF Skull Point Cloud

ceramic pot from Paynes Creek salt works, possibly used for ancient Maya salt production

Jan 28 Pot Obj Post-meshlab2

Exploring photoscan techniques with Visual SFM, CMVS and Meshlab.
Those are very low poly models obtained after importing the photoscans highres models  in Blender. Then I made a quick  retopology and texture baking.

Suitcase, CDs and lamp

Scanned with a DAVID SLS2, 12 images at 30° rotation.
Fully automatic fusion.

Postprocessing: Factor 4 triangle reduction using meshlab to make it fit into 50MB max upload size.

Done at FabLab Nürnberg with https://wiki.fablab-nuernberg.de/w/David_SLS2

The original model is a bronze statue full body, 40cm high.

Asset Library

About These Assets