VISION AWARD
Top vision innovations shortlisted for Stuttgart award
Technology that advances 3D imaging, makes lenses more resistant to vibration, turns a CMOS camera virtually into a CCD, and makes SWIR imaging less expensive, are all innovations that have been shortlisted for this year’s Vision Award, to be presented at the Vision show in Stuttgart, Germany. Imaging and Machine Vision Europe is proud to sponsor the award, which highlights outstanding innovations in machine vision. Here are the seven shortlisted entries
Inline Computational Imaging: single sensor technology for simultaneous 2D and 3D high definition inline inspection By Svorad Stolc, Petra Thanner, and Markus Clabian, Austrian Institute of Technology
T
he Austrian Institute of Technology’s (AIT) Inline Computational Imaging (ICI) platform is an industrial inline
inspection system based on a single camera, which is capable of simultaneous colour 2D and 3D inspection. Te solution combines light field and photometric stereo imaging techniques, and is developed for high- performance industrial inline inspection. Light field and photometric stereo are
computational imaging methods using several images of an object to reconstruct it in 3D. Te light field technique takes multiple images from different observation angles to give good absolute depth accuracy, while photometric stereo makes use of different illumination angles to image fine surface details. Te AIT ICI system combines the advantages of both techniques, together with line scan imaging. A matrix camera fitted with a standard non-
telecentric lens and two static light sources is used in a multi-line scan mode, with only a small number of line-pairs read out while the object moves underneath the camera. Each of these line-pairs acts as an individual line scan camera, producing an image stack containing
All algorithms are fast and accurate enough for high-performance inline inspection
light field information stemming from different viewing angles, and photometric information stemming from different illumination angles. Te ICI 3D reconstruction algorithms are
a combination of depth analysis and surface reconstruction. For the light field depth analysis, the acquisitions are represented as an epipolar plane image (EPI) stack. Te slopes (angles) of the EPI linear structures correspond to the depths of each image position. Depth analysis is performed using a multi-view correspondence analysis in the EPI domain with tailored image features that are insensitive to brightness and contrast variations. In the photometric stereo analysis,
surface orientations are derived from local shading information. Although, the surface geometry can be partly retrieved from surface orientations only, the ICI method combines the depth estimation from the light field and photometric stereo analyses to achieve improved reconstruction accuracy. Final 3D reconstruction is delivered as a point cloud, together with measurement confidence values for each single captured point. Besides the 3D point cloud, pixel-rectified colour images are also captured. AIT ICI comes with a tailored bundle
adjustment algorithm for inline compensation of motion artefacts. Tis makes the solution very robust and allows for high magnification
34 Imaging and Machine Vision Europe • October/November 2018
down to 4μm per pixel lateral resolution. Typically, the system runs at 20kHz at 11
lines. Multi-line scan rates up to 150kHz are feasible using four lines and a high-speed camera by AIT. Tis makes ICI suitable for a broad range of industrial inspection tasks. All algorithms are fast and accurate enough for high-performance inline inspection. All outputs are provided in standard image formats so they can be processed using standard machine vision libraries. Te AIT ICI technology aims to close the
gap between 3D machine vision systems offering either high optical resolutions or high inspection speeds. Te AIT ICI is a rather simple, yet powerful, acquisition concept. In combination with smart algorithms it enables a new dimension in fast and accurate inline inspection.
www.ait.ac.at/en/research-fields/ high-performance-vision/ inline-computational-imaging
@imveurope
www.imveurope.com
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72
