Categories

Archive

Disclaimer

De meningen ge-uit door medewerkers en studenten van de TU Delft en de commentaren die zijn gegeven reflecteren niet perse de mening(en) van de TU Delft. De TU Delft is dan ook niet verantwoordelijk voor de inhoud van hetgeen op de TU Delft weblogs zichtbaar is. Wel vindt de TU Delft het belangrijk - en ook waarde toevoegend - dat medewerkers en studenten op deze, door de TU Delft gefaciliteerde, omgeving hun mening kunnen geven.

DICOM to STL Breakthrough!

After spending many hours on the internet, we finally managed to convert some of the data of the scanned test glass. The scan of the glass (that was done by the smaller Nanoton) has produced several slices. Because of the high resolution and high accuracy of the scanner it overwhelmed us with well over 1000 slices, resulting in a 3GB DICOM file. Converting this data to a 3D-printer extension (like STL) is not as simple as it sounds. The slices are horizontal scans, pictures you might say, made every fraction of a millimeter. These pictures do not contain any data that represent a coordinate system. Therefore, boundary edges for every slice have to be made and a computer program that can combine these edges has to be used. When this has been done, this program will combine every boundary edge and shape them into the final model. This model (now consisting out of points rather than image data) can be exported in STL format. This can be loaded into a modeling application, like Rhinoceros 4. Again, this is more difficult as it sounds. The 3GB of data, or 1152 slices, is too much for a home pc to handle. Therefore, sections of maybe a 100 slices at a time have to be made. Yesterday, as a test, we managed to successfully convert one single slice into a STL file and loaded this Rhinoceros 4. The result is shown below. Our next step is to try and combine a set number of segments into a complete model.

 

Be Sociable, Share!

2 comments

Thanks for your contribution Gerwin de Haan! I had a look through the paper and found the used method for conversion far too advanced for the accuracy we are working with. In the end we got it working with an application that could extract the point clouds from the scan data. This we used to reconstruct the new model. Thanks for reading this blog!

Some of my (former) colleagues collaborated with Material Engineering reconstructing scanned (synchrotron) artefacts using large size dataprocessing and 3D printing:

http://graphics.tudelft.nl/Projects/PrayerBead

this paper contains some software references:

http://graphics.tudelft.nl/Publications/Reischig2009

good luck!

Leave a Reply

© 2011 TU Delft