3D machine vision blog - Zivid

Zivid SDK v1.6

Written by Øyvind Borgan | 2019-11-07

3D hand-eye calibration and fast native 2D imaging.

At Zivid we are committed to continued advancements of the Zivid One+ 3D camera family through regular software updates.

For the release of version 1.6 of the Zivid SDK, our primary focus is to make it easier to increase the performance of our 3D cameras. 

3D hand-eye calibration

We now provide a 3D based hand-eye calibration for Zivid 3D cameras. This solution provides significantly better accuracy than comparable solutions using 2D based calibration. 
 

Hand-eye calibration describes the transformation required to make a 3D camera and a robot work together successfully in one common coordinate system.

Two typical system configurations when performing Zivid 3D hand-eye calibration. The 3D camera can move with the robot arm, or capture the scene from a fixed (static) position relative to the robot arm.

The following two graphs show typical rotational- and translation errors as a function of the number of images used per calibration.

 
Zivid SDK 1.6 includes 3D calibration solutions outperforming comparable methods.

We believe using the calibration patterns and hand-eye calibration provided by the Zivid SDK drastically simplifies the process, and will significantly improve the transformation between the vision component and the robot arm's end-effector.

Provided with SDK 1.6 you will find

  • functions for performing 3D hand-eye calibration
  • hand-eye sample code
  • the required checkerboards as printable pdf files
  • step-by-step guides in the knowledge base.

Visit the Zivid Knowledge Base for the in-depth resources.

Fast native 2D imaging

We have added new functionality for directly capturing 2D images from Zivid One+ 3D cameras.

Note: Zivid One+ uses the same 2.3 MP (megapixel) sensor to capture 2D images and 3D data. This means that you always get a 1:1, pixel-by-pixel, correspondence between 2D and point cloud (3D).

Using this function makes it possible to capture a single 2D image without capturing and processing additional 3D data. This can be used to reduce cycle times when using 2D data in your application.

Save capture settings in Zivid Studio

You can now save (export) each frame's capture settings from Zivid Studio.

In Zivid Studio you can now save each frame's capture settings.

GenICam for HALCON in Linux

Zivid's implementation of GenICam for HALCON is now supported in both Windows and Linux environments.

Additional improvements

In addition to new hand-eye calibration methods and 2D api, we made some under the hood SDK enhancements like improved Open CL diagnostics.
View full post