4 ms·

An End-to-End Guide to Photogrammetry with Mobile Devices

77 points by sebg 5y ago

q_andrew 5y ago
As a 3D artist, photogrammetry is one of my favorite casual creative activities. Compared to the time it would take to model these environments, it almost feels like cheating. I recommend it to anyone with enough computing power!
Youden 5y ago
If you don't need publication-quality 3D models, the latest iPhones have built-in LIDAR sensors, both front (for facial recognition) and back (for AR). There are a variety of apps that can use these built-in sensors to build 3D models of the environment in what I think is a pretty painless and intuitive way. The results aren't publication-quality like the NYTimes' guide's results but it's quick, easy and "good enough" for many purposes. Personally, my preferred scanning app is Heges [0]. Record3D [1] by the same developer is also really cool, albeit not a scanner app. One note though: if you want high-quality 3D geometry, use the front LIDAR sensor, not the back. The front sensor is much more precise. [0]: https://hege.sh/ https://hege.sh/ [1]: https://record3d.app/ https://record3d.app/
- cyberge99 5y ago
  Along those lines, I use Scaniverse. I believe they were recently acquired by Nanant maker of Pokemon Go. Their vision (no pun intended) is to scan the whole world Pokemon Go style. An interesting endeavor if you ask me.
  - easton 5y ago
    I wonder why their app doesn't support the front-facing LIDAR sensor like the others. (Also, mandatory "Google bought it I bet it's shut down in 6 months". There, now nobody else has to say it.)
stjo 5y ago
At the startup I work at we make a platform allowing users (mainly telecom providers) to upload drone images from some site. Then we generate a 3d model and do additional computer vision tasks such as finding rusty parts, simulate coverage to find more optimal antenna placements, real time collaboration etc. Have a look at: https://platform.myxrobotics.com/twins/109/?invcode=Db4GksYM https://platform.myxrobotics.com/twins/109/?invcode=Db4GksYM https://platform.myxrobotics.com/twins/3/?invcode=EhiGcTX- https://platform.myxrobotics.com/twins/3/?invcode=EhiGcTX- The technology stack is pretty similar to nytimes's: 1) Compute the model and divide it into 3d tiles. 2) Load the model in the browser and overlay changes on top. Except that instead of story telling we help telecommunication engineers :D

77 points by sebg 5y ago

As a 3D artist, photogrammetry is one of my favorite casual creative activities. Compared to the time it would take to model these environments, it almost feels like cheating. I recommend it to anyone with enough computing power!

Youden 5y ago

If you don't need publication-quality 3D models, the latest iPhones have built-in LIDAR sensors, both front (for facial recognition) and back (for AR). There are a variety of apps that can use these built-in sensors to build 3D models of the environment in what I think is a pretty painless and intuitive way. The results aren't publication-quality like the NYTimes' guide's results but it's quick, easy and "good enough" for many purposes. Personally, my preferred scanning app is Heges [0]. Record3D [1] by the same developer is also really cool, albeit not a scanner app. One note though: if you want high-quality 3D geometry, use the front LIDAR sensor, not the back. The front sensor is much more precise. [0]: https://hege.sh/ https://hege.sh/ [1]: https://record3d.app/ https://record3d.app/

cyberge99 5y ago
Along those lines, I use Scaniverse. I believe they were recently acquired by Nanant maker of Pokemon Go. Their vision (no pun intended) is to scan the whole world Pokemon Go style. An interesting endeavor if you ask me.
- easton 5y ago
  I wonder why their app doesn't support the front-facing LIDAR sensor like the others. (Also, mandatory "Google bought it I bet it's shut down in 6 months". There, now nobody else has to say it.)

stjo 5y ago

At the startup I work at we make a platform allowing users (mainly telecom providers) to upload drone images from some site. Then we generate a 3d model and do additional computer vision tasks such as finding rusty parts, simulate coverage to find more optimal antenna placements, real time collaboration etc. Have a look at: https://platform.myxrobotics.com/twins/109/?invcode=Db4GksYM https://platform.myxrobotics.com/twins/109/?invcode=Db4GksYM https://platform.myxrobotics.com/twins/3/?invcode=EhiGcTX- https://platform.myxrobotics.com/twins/3/?invcode=EhiGcTX- The technology stack is pretty similar to nytimes's: 1) Compute the model and divide it into 3d tiles. 2) Load the model in the browser and overlay changes on top. Except that instead of story telling we help telecommunication engineers :D