OmniTek OTR 1001, Руководство пользователя

Страница: 126 / 390

OmniTek OTR 1001 User Guide, v3.1 I-2
Depth The distance a 3D object appears to be from the screen plane. Positive values
correspond to positions behind the screen; negative values correspond to
positions in front of the screen.
Depth Budget The desired range of depths within which all 3D objects should appear, expressed
either in terms of a maximum (most positive) and a minimum (most negative)
disparity or a maximum (most positive) and a minimum (most negative) depth.
I.1 Stereo 3D Overview
Stereo 3D video produces the effect of 3D by deploying the brain’s capacity to interpret parallax
between right- and left-eye images of the same object as a single object at a distance from the viewer.
For the viewer’s brain to merge two images into one 3D image:
1. The ‘disparity’ (see above) must lie between the value representing the closest object
(typically no less than about 1m in front of the viewer) and the value representing the
farthest object (usually taken to be at infinity)
2. The colorimetry of the two images must match.
The OTR offers a variety of ways to display the depth information that can be deduced from the
two signals, together with a range of ways to compare the colour information in the two
component video signals. It also offers a choice of Picture modes through which to see the effect
of combining the images.
Other facilities include automatic read-out of the coordinates in the viewer’s world of any point you
pick with the crosshair cursor (i.e. the
x , y , and z coordinates that the selected item will appear to
have relative to the viewer) and the ability to limit the area being analysed to a specified Region of
Interest.
The crosshair cursor and/or its coordinates in the viewer’s world are shown on many of the displays
described in this section. This information also appears in the Properties control. Any change in the
position of the crosshair cursor in any of these displays automatically causes all references to this
cursor to be updated to reflect the new position in the viewer’s world coordinates. The Regions of
Interest shown on the left- and right-eye images are similarly linked by an offset corresponding to the
measured disparity.
The right- and left-eye images are typically provided as separate inputs to the OTR, but may also be
delivered Side-by-Side (SbS) or Top + Bottom (TaB) or taken from 3G Level B Dual-Stream video.
Where separate inputs are used, the user specifies which input is providing the left-eye image and
which the right-eye image. The thumbnails on the Status Bar automatically switch to show the left-
eye image to the left and the right-eye image to the right. Where the video is taken from a mirror rig,
the user specifies – for both Left and Right image streams – whether the image is in the normal
orientation, flipped horizontally or flipped vertically.
The use of a mirror rig typically also means that the colours are generally less bright in the image that
has been through the mirror. Such differences in colour adversely affect the OTR’s ability to match
objects across the left-eye and right-eye images. The process of matching objects is also adversely
affected by any vertical disparity between the images. However, the OTR is able to determine the
corrections that need to be applied in each case, and ‘Auto gain control’ and ‘Auto correct vertical
disparity’ options are offered. When selected, these options can significantly improve both the
quality of the 3D depth analysis and the OTR’s assessment of camera set-up issues such as
differences in zoom and camera roll.
Important: Processing Stereo 3D video takes significant resources. In particular, it is not
possible to both analyse Stereo 3D and to capture images at the same time. The OTR therefore
allows you to enable and disable 3D monitoring as required: this is described in Section I.2.