ReconstructMe SDK has now an open source user interface. It is called ReconstructMeQt and provides intuitive reconstruction experience. ReconstructMeQt is an open source project available at Google Code.
We invite interested developers to extend the project.
As of today ReconstructMe SDK 1.0 is available for commercial and non-commercial use. ReconstructMe SDK is designed for a broad band of applications. It targets simple real-time reconstruction applications and scales up to multiple sensor projects.
We spent a lot of time in designing the API as we wanted it to be as easy-to-use as possible without limiting the possible flexibility. After dozens of approaches and evolutionary prototypes, we’ve settled with a design that we think accomplishes the following main goals
Real-time 3D reconstruction is complex. We tried to hide as many details as possible about the process to allow you to concentrate on application programming.
The ReconstructMe SDK is available for non-commercial usage (with some limitations) and commercial usage. Both feature the same functionality, but the non-commercial version is limited in some aspects. See our project page for details. The package comes bundled with an installer that allows you to install the necessary sensor drivers on the fly (note that the drivers are not included, but downloaded remotely from our site).
ReconstructMe SDK is currently compiled for Windows 32bit using Visual Studio 10.
Getting reconstructme to work in other programming languages is easy. We added an example to demonstrate the case for C#. We’d like to add binding for all common languages, but obviously that is a task that requires the help of the community. If you’d like to contribute to a binding drop us a note in our development forum.
Here’s a very first introductory example into the C-API copied from the reference documentation. The example shows how to perform real-time reconstruction using a single sensor. Besides, built-in real-time visualization and surface visualization is used. The final mesh is saved in PLY format.
Here’s the corresponding code
1 | // Include the necessary headers |
2 | #include <reconstructmesdk/reme.h> |
1 | // Create a new context |
2 | reme_context_t c; |
3 | reme_context_create(&c); |
4 |
5 | // Compile for OpenCL device using defaults |
6 | reme_context_compile(c); |
7 |
8 | // Create a new volume |
9 | reme_volume_t v; |
10 | reme_volume_create(c, &v); |
11 |
12 | // Create a new sensor. Tries multiple backends using default |
13 | // sensor configurations, returns the first working one. By default |
14 | // each sensor works on the first volume created. |
15 | reme_sensor_t s; |
16 | reme_sensor_create(c, "openni;mskinect;file", true, &s); |
17 | reme_sensor_open(c, s); |
18 |
19 | // For debugging purposes open a viewer for tracking the reconstruction process. |
20 | // This will create a two-sided viewer showing the video stream and the current |
21 | // reconstruction from the sensor point of view. |
22 | reme_viewer_t viewer; |
23 | reme_viewer_create_image(c, "This is ReconstructMe SDK", &viewer); |
24 | reme_viewer_add_image(c, viewer, s, REME_IMAGE_DEPTH); |
25 | reme_viewer_add_image(c, viewer, s, REME_IMAGE_VOLUME); |
26 |
27 | // Perform reconstruction until no more frames are left |
28 | int time = 0; |
29 | while (time < 200 && REME_SUCCESS(reme_sensor_grab(c, s))) { |
30 |
31 | // Prepare image and depth data |
32 | reme_sensor_prepare_images(c, s); |
33 |
34 | // Try to determine updated sensor position. |
35 | // On succes, update volume, otherwise move to a recovery position |
36 | // and wait for the tracking to start again. |
37 | if (REME_SUCCESS(reme_sensor_track_position(c, s))) { |
38 | |
39 | // Update volume with depth data from the |
40 | // current sensor perspective |
41 | reme_sensor_update_volume(c, s); |
42 |
43 | } |
44 |
45 | // Update the viewer |
46 | reme_viewer_update(c, viewer); |
47 | time += 1; |
48 | } |
49 |
50 | // Close and destroy the sensor, it is not needed anymore |
51 | reme_sensor_close(c, s); |
52 | reme_sensor_destroy(c, &s); |
53 |
54 | // Create a new surface |
55 | reme_surface_t m; |
56 | reme_surface_create(c, &m); |
57 | reme_surface_generate(c, m, v); |
58 | reme_surface_save_to_file(c, m, "test.ply"); |
59 |
60 | // Visualize resulting surface |
61 | reme_viewer_t viewer_surface; |
62 | reme_viewer_create_surface(c, m, "This is ReconstructMeSDK", &viewer_surface); |
63 | reme_viewer_wait(c, viewer_surface); |
64 |
65 | reme_surface_destroy(c, &m); |
66 |
67 | // Make sure to release all memory acquired |
68 | reme_context_destroy(&c); |
69 |
70 | /** [Example Code] */ |
For more examples please visit the documentations’ example page.
Happy reconstruction!
The ReconstructMe Team.
Corridor Digital has created an amazing clip with the help of ReconstructMe.
The sand models are created with a Kinect and ReconstructMe:
ASUS demonstrated ReconstructMe with their sensor Xtion Pro at the Computer Graphics Workshop 2012 helded by ACM SIGGRAPH in Taipei (2012/07/12 – 2012/07/13). Below, you can see some impressions of the workshop.
We have received a note from ASUS that ReconstructMe will be demonstrated live at the Computer Graphics Workshop 2012 held by ACM SIGGRAPH in Taipei (2012/07/12 – 2012/07/13).
ASUS will introduce ReMe at their booth. Picture and video footage will be uploaded when available.
Elliot has put a new video tutorial online that shows how to setup re-projection of virtual data on real-world objects using a video projector. He uses ReconstructMe and a software called VVVV.
While making some serious progress on texturing scanned surfaces, we ran into need of a more decent surface reconstruction and decimation technique. Until now, exported meshes contained hundreds of thousands of triangles, adding unnecessary overhead in regions that could be expressed with just a couple of triangles (e.g planar regions). Additionally, we felt the need of closing small surface holes in order to allow smooth texturing across the surface.
Therefore we re-designed our surface reconstruction pipeline to support more sophisticated reconstruction techniques and a configurable surface decimation pipeline.
Below is an image that shows the original mesh as generated by the current version of ReMe (v. 0.6.0-405). It contains roughly 250.000 faces and one can clearly spot the holes that remained due to the lack of visibility of these areas while scanning.
In contrast, the next image shows a successful reconstruction of the original surface reduced to 50.000 faces with boundary holes closed.
Comparing both meshes using the Hausdorff distance gives an average distance of 0.8 mm. The image below colorizes the distances (blue low, red high).
Surface reconstruction isn’t limited to individual meshes, but can also be used to fusion multiple volumes into one single consistent mesh. The image below shows two individual stitched meshes using ReMe’s --multiscan feature.
Here is the fusioned mesh as generated by the development version of ReMe
Stay tuned!
Taken straight from Thingiverse
Mike Moceri and Tom Burtonwood collaborated on a project that premiered May 26th at The Southside Hub of Production for the exhibition “On Making Things Matter”. We 3d scanned (with the Kinect, Reconstructme + Netfabb) visitors at the opening reception and then 3d Printed them with a MakerBot Thing-O-Matic. All the “portraits” we scanned are uploaded to Thingiverse to share with teh internets. We have also scanned portions of the exhibition and the interior of the building and we will be installing 3d prints of these vignettes over the course of the exhibition.
Here’s a video (you can see ReMe in action at the beginning of the video)
Mike attended the Zhou B. Art Center in Chicago participating in an event called Facemask. More on this including a video stream can be found here.
MagWeb dropped us a note about his latest reconstructions of altars using the new multi-volume stitching method. He calls Reme the “by far the fastest [scanning] system for big stuff!”. Here are three reconstructions made by multiple 300cm volumes. Enjoy!
Here are two wire-frame zoom images
We have just launched our commercial version which can be used for applications where monetary remuneration takes place. Commercial licenses are available at our web shop.
Please consider to support us!
Additionally we’ve updated ReconstructMe to 0.6.0-405. Here’s the changelog
Thanks to all that contributed in our feature survey on Google+. We adapted our internal road-map due to it’s results and will head for texture support and an easy to use graphical user interface next.
Even though it’s still early days, we’d like to share what can be accomplished using our texturing engine. We are aiming at a semi-automatic solution where users are attaching textures through a post-processing step. This allows us to use arbitrary cameras for texturing the mesh. Our engine supports multiple textures and combines them smoothly.
The first image below shows the plain surface reconstructed in high resolution mode. Next to it an image taken with an Casio Exilim ZR 100 to be used as texture. On the right the final textured mesh is shown.
The best part about it, it took less than 3 minutes from scanning to texturing for the above scenario. We will probably release this feature alongside with the first version of the UI around late summer.
Enjoy!
This is worth sharing! Tony has gathered a lot of experience in generating customized PEZ dispensers. It all started with a single dispenser for his own use and he ended up producing 92 PEZ heads till this day. A chronology in pictures.
More information on his photostream.
Images are licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License.
This is a press release in German language.
MÜNCHEN / STEYR. „Next Level“ – Fabrikshalle: Wenn die Softwareingenieure von PROFACTOR die Kamera „Kinect“ der Microsoft Spielkonsole Xbox 360 in die Hand nehmen, haben sie keinen neuen Punkterekord im Auge. Das „Spiel“ beim österreichischen Forschungsunternehmen heißt ReconstructMe und zielt auf Anwendungen in der Industrie ab: „Augen“ für den Roboter und das blitzschnelle Erstellen von CAD-Modellen oder Arbeitsumgebungen werden damit Realität.
Die Kamera „Kinect“ der Spielkonsole Xbox 360 ist in der Lage, Bewegungen von Spielern vor dem Fernseher zu detektieren. PROFACTOR nutzt das System zur 3D-Objekterkennung.
Dazu wurde von PROFACTOR Software-Ingenieuren ein neuer Algorithmus implementiert. Er ermöglicht den Einsatz der Kamera als Scanner. Dieser beschränkt sich nicht auf die Aufnahme von Bildern, sondern erkennt Objekte – und das in Echtzeit. Möglich ist das durch die Verknüpfung der von der Kamera aufgenommenen Bilder mit dem Bewegungspfad der Kamera.
Das exakte Modell eines beliebigen Gegenstandes oder einer Arbeitsumgebung erfolgt in Sekundenschnelle. Derzeit lotet PROFACTOR aus, wie das System in der industriellen Praxis eingesetzt werden kann.
ReconstructMe – Einsatzmöglichkeiten
PROFACTOR auf der Automatica in München: Halle B2 Stand 229
What’s the next most important feature to be implemented in ReconstructMe?
Please join our survey at Google+. We need your opinions.
