[example/device/framegrabber/saveRealSenseData] Wrong camera parameters and depth_M_color homogeneous matrix when aligned depth is requested

[JuInria]

System

• Ubuntu 20.04
• Realsense 2.50.0 (built from source)
• ViSP v3.5.0 (built from source)
• Intel D435

./example/device/framegrabber/saveRealSenseData

When aligned depth is requested -a:

the depth and the color camera parameters should be similar: ROS-based example:

Here, with saveRealSenseData, the resulting camera parameters are: camera.xml:

<?xml version="1.0"?>
<root>
  <!--This file stores intrinsic camera parameters used
   in the vpCameraParameters Class of ViSP available
   at https://visp.inria.fr/download/ .
   It can be read with the parse method of
   the vpXmlParserCamera class.-->
  <camera>
    <!--Name of the camera-->
    <name>color_camera</name>
    <!--Size of the image on which camera calibration was performed-->
    <image_width>640</image_width>
    <image_height>480</image_height>
    <!--Intrinsic camera parameters computed for each projection model-->
    <model>
      <!--Projection model type-->
      <type>perspectiveProjWithDistortion</type>
      <!--Pixel ratio-->
      <px>605.146728515625</px>
      <py>604.79150390625</py>
      <!--Principal point-->
      <u0>325.53253173828125</u0>
      <v0>244.95083618164062</v0>
      <!--Undistorted to distorted distortion parameter-->
      <kud>-0</kud>
      <!--Distorted to undistorted distortion parameter-->
      <kdu>0</kdu>
    </model>
  </camera>
  <camera>
    <!--Name of the camera-->
    <name>depth_camera</name>
    <!--Size of the image on which camera calibration was performed-->
    <image_width>640</image_width>
    <image_height>480</image_height>
    <!--Intrinsic camera parameters computed for each projection model-->
    <model>
      <!--Projection model type-->
      <type>perspectiveProjWithDistortion</type>
      <!--Pixel ratio-->
      <px>381.7528076171875</px>
      <py>381.7528076171875</py>
      <!--Principal point-->
      <u0>323.32614135742188</u0>
      <v0>236.82505798339844</v0>
      <!--Undistorted to distorted distortion parameter-->
      <kud>-0</kud>
      <!--Distorted to undistorted distortion parameter-->
      <kdu>0</kdu>
    </model>
  </camera>
  <camera>
    <!--Name of the camera-->
    <name>infrared_camera</name>
    <!--Size of the image on which camera calibration was performed-->
    <image_width>640</image_width>
    <image_height>480</image_height>
    <!--Intrinsic camera parameters computed for each projection model-->
    <model>
      <!--Projection model type-->
      <type>perspectiveProjWithDistortion</type>
      <!--Pixel ratio-->
      <px>381.7528076171875</px>
      <py>381.7528076171875</py>
      <!--Principal point-->
      <u0>323.32614135742188</u0>
      <v0>236.82505798339844</v0>
      <!--Undistorted to distorted distortion parameter-->
      <kud>-0</kud>
      <!--Distorted to undistorted distortion parameter-->
      <kdu>0</kdu>
    </model>
  </camera>
</root>

Furthermore, in this case (depth aligned to the color frame), depth_M_color should be an identity matrix. With or without the -a option, we're getting the same homogeneous matrix. In my case:

0.999972403  -0.00671335822  -0.003179509426  -0.01465611346
0.006699616555  0.9999682307  -0.004313052632  9.024870815e-05
0.003208363429  0.004291632213  0.9999856353  -0.0004482123477
0  0  0  1