View allAll Photos Tagged MeshLab
There's background on the B-23 and how it crashed here:
www.flickr.com/photos/flickrdave/3925755985
As for the image itself, here's the story. Back in 2009 I attended an aviation archaeology field school. (Long, stupid story....) Anyway, I had my pole aerial photography gear with me and a couple of us thought a good way to survey and document the site would be to take overhead shots of everything. We were hoping to stitch them all together to make a photo-realistic map of the site. The images we got were interesting but, unfortunately, we weren't able to combine all the images into the map we wanted. One software package I had heard of was too expensive for us and another that I did get my hands on was too fussy and wouldn't work with our raw images.
A few days ago, though, I learned about a new web site called Hypr3d.com that lets you upload images and it creates 3D models from them. I didn't mess around. I uploaded all 145 clear images we had of this site and let it chew on that overnight. By morning it had produced a rough model I could view in my browser and a higher resolution one that I could download to my PC. I used another application, Meshlab, to rotate it around to the views I wanted and had it render these two images. Both of these views are rendered in "Ortho" mode so there is no perspective. I think that is why it looks less like a photo and more like a painting or model. So, two years after field school ended, I finally have my orthographic projection of the whole site.
In retrospect, I would do this a little differently if I went back. Because all of our images were shot straight down, more or less, there isn't enough information about the vertical faces of the wreck. This shows up as a lot of distorted shapes on the front and sides. We should have taken more oblique shots from the perimeter looking inward. But when you look at the results from an overhead point of view, it looks damn near perfect.
Original 3d scan taken from hyper3d.com, a free photo reconstruction service. Here the scan is previewed in meshlab. Also free.
--------------------------------------
tSet();
tPot(); tCup();
A tea set is created by sampling and remixing existing objects using 3D scanning from photographs (via hypr3D.com) and a custom made Processing sketch that works as an interface to manipulate meshes and create ready-to-print files.
The process is an alternative partcipatory approach to product consumption that exploits emerging technologies to make it possible for people to copy, modify and fabricate objects to their own taste.
A project developed during the Generator.x 3.0 workshop organised by Marius Watz at iMal.
Creating an STL file from Processing with Modelbuilder using logic from Kyle McDonald's Makerbot app: github.com/kylemcdonald/Makerbot
Still need to clean up the edges of the model and then I'll have something solid.
Made with Processing and Marius Watz's Modelbuilding library. Code still a problem with certain vertices ending up zero-ed out (hence the tangle of triangles pulled forward to the origin). In-progress code is here: gist.github.com/1151193
working on legibility by separating letter cells from the neighbourhood. It's a bit too drastic though and I want the effect to look less like simple extrusion. Currently also hitting a boundary in MeshLab which crashes when trying to import models with 4 million+ triangles... :(
A ruined broch, surviving as a mound of rubble, this broch is noted for its outer defences consisting of two ramparts and a wide ditch between them.
Brochs are circular defended houses with thick (and often hollow) walls which contain galleries, stairways, cells and with a guard cell to the side of the entrance passageway. The best surviving broch is on the Isle of Mousa and it stands to about 13 metres high, but is unlikely to be representative of the height of all brochs. In terms of date, excavated brochs have have fallen within the date range of 600BC-100AD (Iron Age) for their construction, although it has to be remembered that they are often part of complex multi period sites. Brochs are mostly found in the north and west of Scotland.
Here at Dalsetter structures and enclosures relating to a much later farmstead also occupy the broch site - the buildings were depicted as being unroofed on the Ordnance Survey 25 inch to the mile 1st edition (Surveyed: 1878 Published: 1882) map (Shetland Sheet LXV.15).
In the top of the broch mound there is a depression containing a small rectangular single cell structure. A short length of facing stones of the presumed outer face of the broch are visible on the south side of the mound, following a similar curvilinear alignment to these stones is a green depression, presumably a robber trench where the rubble mound has been dug into to remove the facing stones for reuse. The limit of the evidence on the site suggests that the external diameter of the broch was upto about 15 metres.
Grid Reference: HU 40754 15662
see: canmore.rcahms.gov.uk/en/site/899/details/dalsetter/
This image is of an artificially lit 3d model the site of the Broch of Dalsetter. The 3D Model was produced using Agisoft Photoscan from 171 kite aerial photographs, which produced a model with 102,375,504 faces and 51,194,033 vertices which I reduced to 20,000,000 faces to do the artificial lighting in Meshlab. A little bit of mathematics tells me the camera was at 45 metres.
3D model derived from Kite Aerial Photographs
17 September 2014
I figured i should go ahead and make a new icon for my signature. What with the new year and new computer and all... Besides, now that I am more familiar with TopMod, I was able to do more along the lines of what my original idea.
This image took 23:26 to render, but with as many sphere lights as it has, plus 32 path samples and aa 3 3, it would have taken well over a week on the old computer...
There were a few holes that had to fill in using MeshLab, and the interface was very easy to use...
TopMod / MeshLab / Structure Synth / Sunflow
I recently got a set of Klixx to fiddle around with at work while my code is compiling, and was inspired to try and re-create them.
Sketchup / Meshlab / Structure Synth / Sunflow
Model of a wearable piece made from 12 months of Canberra weather data, July 08 -June 09. The outer edge is daily temperature data; the holes are weekly aggregated rainfall. Made with Processing, Blender (for the boolean subtraction) and Meshlab. Pardon the cheezy background - a Meshlab snapshot.
TopMod / Sunflow
This is a render of the same object as before, but with the constant-shaded lights cranked up to be much brighter, and the ibl lighting turned off to get more interesting shadows on the ground. Thanks to Fractured Pixel for showing me how to re-triangulate my .obj files by simply opening them in MeshLab and re-saving them with Normals unchecked. This fixes the 'holes' that were present in the previous render.
A basic 3D model created by using two kite aerial images into 123DCatch and then transfered into Maya to create a high density mesh. The model has a basic surface added to it and a simple light model animated across its surface to highlight any potential features. The KAP images can then be replicated in the model and examined for further potential from different viewpoints.
I like the architectural quality of this example. Could imagine this as nice stepping stones/foot bridge semi-immersed in water...
Refactored the IsoSurface class so it's not any longer constrained to regular cubes only, but can use different lengths in each dimension. Also added a feature to close/seal off all sides so the mesh has no holes on the sides and is ready for printing/fabbing...
Shells designed in Processing w/ Hemesh + Meshlab. 3D Printed using Zprinter by 3D Sytems (Zcorp).
Project Title: Listening to the Ocean on a Shore of Gypsum Sand
By:
Gene Kogan (programming)
Phillip Stearns (concept and design)
Dan Tesene (fabrication and concept)
Shells designed in Processing w/ Hemesh + Meshlab. 3D Printed using Zprinter by 3D Sytems (Zcorp).
Project Title: Listening to the Ocean on a Shore of Gypsum Sand
By:
Gene Kogan (programming)
Phillip Stearns (concept and design)
Dan Tesene (fabrication and concept)
using mirrors to turn a single kinect into 5 kinects, scanning all sides of an object simultaneously.
this scan represents the average of a short 10 second video while moving a small aperture in front of the camera to avoid interference between reflections of the projected ir pattern.
TMR posited it might take half an hour to make a d20 in Sketchup... it took 35-40 minutes, including figuring out a good process.
For the record, I used TopMod to generate the original twenty-face regular solid (it\'s an icosahedron, I believe, but my memory may be faulty), and I also cleaned up in Meshlab (couple clicks just to make sure the mesh was ok).
It Skeined great, at least for the jumbo size dice.
Collection of the different mask renders + lighting models used for the poster/catalog image. The image was then assembled from these in Photoshop...
The original mesh is based on a 128x128x128 volume.
meshLab is a great tool for previewing STL files. It's fast, good-looking and gives you immediate visual feedback on whether your surface normals are correct or not.
meshLab is a great tool for previewing STL files. It's fast, good-looking and gives you immediate visual feedback on whether your surface normals are correct or not.
It also has some rather less useful but fun shaders, such as this polka dot shader.
A basic 3D model created by entering some of my older images into a newer version of Photoscan. The presented meshes appear much more detailed than previously created and therefore show potentially more features. The generated dense point cloud is converted into a high resolution mesh and imported into MeshLab to manipulate lighting and shading. The KAP images can then be replicated in the model and examined for further potential from different viewpoints.
A basic 3D model created by entering some of my older images into a newer version of Photoscan. The presented meshes appear much more detailed than previously created and therefore show potentially more features. The generated dense point cloud is converted into a high resolution mesh and imported into MeshLab to manipulate lighting and shading. The KAP images can then be replicated in the model and examined for further potential from different viewpoints.
You can clearly see the outer defensive ramparts and the raised area within.
Shells designed in Processing w/ Hemesh + Meshlab. 3D Printed using Zprinter by 3D Sytems (Zcorp).
Project Title: Listening to the Ocean on a Shore of Gypsum Sand
By:
Gene Kogan (programming)
Phillip Stearns (concept and design)
Dan Tesene (fabrication and concept)
Shells designed in Processing w/ Hemesh + Meshlab. 3D Printed using Zprinter by 3D Sytems (Zcorp).
Project Title: Listening to the Ocean on a Shore of Gypsum Sand
By:
Gene Kogan (programming)
Phillip Stearns (concept and design)
Dan Tesene (fabrication and concept)
First sketches showing the potential of the new volumetric brush (here the box version was used). Size, density and brush mode (additve, multiply or replace) can be customized. Here I first used a massive brush size with high positive density to create the box, then switched to negative density and smaller size to carve out. The final mesh has 4x Laplacian smooth applied and the colours are visualizing curvature and were also applied in Meshlab later.