Maciej Karaszewski, Robert Sitnik

Size: px

Start display at page:

Download "Maciej Karaszewski, Robert Sitnik"

Bryan McDonald
5 years ago
Views:

1 European Cooperation in Science and Technology Towards the automation and objectivity of 3D scanning Maciej Karaszewski, Robert Sitnik Warsaw University of Technology, Faculty of Mechatronics

2 Automation what for? High resolution digitization (structured light, photogrammetry, structure from motion etc.) High spatial resolution (ex pts / mm 2 ) Small working volume: 4000 pts / 100(pts/mm) = 40 mm Usually small depth (10 20 mm) Many measurements required for digitization (up to few thousands) Limited resolution of detector (ex. 16MPx, square, 4000x4000 px)

3 4451 directional measurements 10 mm 1 mm 1 mm Object s size: 500 mm radius / 2000 mm height Resolution 0.02 mm (2500 points / mm 2 ) (measurement volume 55mm x 55mm x 20 mm) 1 mm

4 Automation what for? Objectivity and repeatability Operator s attributes: Experience Patience Precision Carefullness Operator s disposition: Mood Fatigue Health External paramters like weather, humidity, etc. General limitations: Limited repeatability of manual operations Performance drop due to long-lasting similar activities Accidents and carelessness Influence of temperature, humidity etc. Digitization results of the same object in the similar condition are not repeatable!

5 What to automate? Manipulation of scanning head Image courtesy of: Hexagon Metrology Cognitens WLS400M Image courtesy of: RangeVision RVScanner Image courtesy of: Creaform Go!Scan 3D scanner

6 What to automate? Planning next measurements

7 What to automate? Initial and fine integration of measurements Image courtesy of: Technabob Image courtesy of: Creaform Image courtesy of: Research.history.org Image courtesy of: Leica Geosystems

8 Our solution M. Karaszewski, R. Sitnik, E. Bunsch, On-line, collision-free positioning of a scanner during fully automated three-dimensional measurement of cultural heritage objects, Robotics and Autonomous Systems, Volume 60, Issue 9, September 2012, Pages

9 Why so complicated? Occlusions Various surface orientation Size Small depth of measurement volume Image courtesy of: Solutionix Rexscan CS+ 3D

10 Calibration

11 Start of measurement

12 Next best view

27 Collision avoidance

28 Data processing Automatic noise filtering Automatic initial integration ICP algorithm for fine integration Simplification algorithms, textured triangle mesh generation Automatic preparation for real-time visualization

automated threedimensional measurement of cultural heritage objects,

29 Exemplary objects From: M. Karaszewski, R. Sitnik, E. Bunsch, On-line, collision-free positioning of a scanner during fully automated threedimensional measurement of cultural heritage objects, Robotics and Autonomous Systems, Volume 60, Issue 9, September 2012, Pages

30 Exemplary objects Times in this table are given for the complete digitization process. About 60-70% of this process is used for data processing which does not require any human supervision.

www.cost.esf.org European Cooperation in Science and Technology Thank you for your attention! M. Karaszewski, R.

31 European Cooperation in Science and Technology Thank you for your attention! M. Karaszewski, R. Sitnik m.karaszewski Warsaw University of Technology, Faculty of Mechatronics

Manufacturing Metrology Team

Manufacturing Metrology Team The Team has a range of state-of-the-art equipment for the measurement of surface texture and form. We are happy to discuss potential measurement issues and collaborative research Manufacturing Metrology