A Bumblebee Stereo Camera mounted on a Pan-Tilt-Head aquires pairs of images, which are transformed to Point Clouds. Multiple Point Clouds are joined to a common Point Cloud.
In this video the internal data structure of the Point Clouds: the Octree is shown. With this tree structure many point cloud operations can be calculated much faster than with a simple list of points.
This
video shows a simple scene with a box on a
table and how it is
seen by a stereo camera and a time-of-flight camera. All three camera
views can be visualized at the same time in the same 3D-environment
In this
video a marker detection of a known object (table plate) is
shown. There is a discrepancy between the 3D-model of the table plate
and the image plane at the far end of the table. In addition, the image
and data inspection capabilities of ImageNet Designer are shown.
This
video shows the integrated help in each ImageNets block which is
displayed when the mouse cursor rests some time over the block. There
is a general help for the whole block, a help for every input and
output and for every property of a block. The author(s) can also be
defined and displayed.
Feedbacks
can be used to improve automatic detection even under
changing illumination conditions. This is especially useful in service
robotic applications.
This
video shows a colored point cloud captured by a stereo camera. The
point cloud is thresholded by a plane which is defined by a coordinate
transformation frame. The frame itself is the output of a marker
detection.
This
video shows the calculation of a color point cloud and that the
same functionality can be loaded in a SubNet Block.
