CVB 2018 release


Hello all,
as you guys are the cognoscenti, the inner circle, you get to know first: the release of CVB 2018 is here. It is already in the :cvb: forum’s downloads section for Windows 32/64, Linux 32/64 and ARM 32/64.

So it is new, but what does it bring to you? Quite a lot. Those of you that came to the CVB User Meeting had a preview of some of it:

  • Image Manager / CameraSuite - we’ve added what we call core3D functionality. This means acquisition of 3D data from a camera, loading and saving of 3D data in various formats and, of course, display. The data can be calibrated or uncalibrated. As well as the SDK functionality, there is also a standalone 3D Player application. This can display 3D images, but also allows the user to crop, filter and apply look-up tables, all without writing any code - a good tool for set-up, proof-of-concept and debugging 3D images. There are other smaller changes too: a new ‘chunk parser’ in the nodemap to allow chunk data from GenICam devices to be accessed more easily. TurboDrive support for Teledyne DALSA cameras has been optimised for performance, reducing the CPU load.
  • Foundation Package - big changes here too. Foundation’s Metric functionality adds 3D calibration to the existing 2D calibration. Using a calibration target allows distortions to be removed as well as moving into real-world co-ordinates. Foundation also gains a new barcode reader called ZXBarcode, with particular strengths in reading QR codes. Some of the existing Foundation tools have also been ported to Linux (which means ARM platforms too) - Arithmetic, Edge, BayerToRGB, Lightmeter and TextOut. Previously we had a separate tool called Optical Flow, this has now been integrated into Foundation. The idea is to make the Foundation Package really excellent value. If you haven’t looked at Foundation before, we are trying to make it a compelling choice.
  • Match 3D - you may remember we had a tool with this name before… This one is a completely new development, which has meant that we have been able to bring it to Linux from the start too. Match 3D allows a target pointcloud and a template pointcloud to be compared to find the misalignment. This allows two main use-cases - robot pick-and-place or subtraction of the two aligned pointclouds to create a difference image, showing height differences in 3D - a useful 3D defect detector.

Over the next few days there will be more detailed information available on the CVB website, news stories and press releases - but you heard it here first…

If you have any questions or comments, please let me know. :man_technologist:

pinned globally #2