Processing camera rigs

So-called camera rigs are systems connecting several cameras by rigs. With one trigger command, referred as a capture, the system enables all the cameras connected in the rig to take images simultaneously (with a time difference much less than the image read-out time, which should be in millisecond-level).

How are the images assigned to the different rig instances?

Parrot Sequoia and MicaSense RedEdge cameras are processed in Pix4Dmapper as camera rigs. For both Parrot Sequoia and MicaSense RedEdge cameras, when a capture is triggered, an image per sensor is taken (i.e. for Parrot Sequoia 1 capture contains 4 images since it has 4 sensors: G, R, RE, NIR; for MicaSense RedEdge 1 capture contains 5 images since it has 5 sensors B, G, R, RE, NIR). These captures are assigned to one rig instance to calibrate the rig successfully.

 
Note: The new rig assignment algorithm with Parrot Sequoia or Micasense Rededge is now available, starting with Pix4D Desktop 4.2.25.

The rig assignment algorithm uses the unique identifier associated with each capture, available from the image EXIF of Parrot Sequoia or Micasense Rededge cameras, to group the images from a capture and assign it to rig instances.

Moreover, in case of missing images per rig instance, the rig instance will be identified as incomplete and further disabled. Only a limited amount of incomplete captures is allowed for the processing to continue. If more than 5% of the total number of captures are incomplete the processing will stop.

How to edit the rig model parameters and processing options?

Access camera rigs parameters

To access camera rigs parameters follow the instructions:

1. On the menu bar, click Project > Image Properties Editor...
2. In the section Selected Camera Model, next to the rig model name, click Edit...
2018-04-05_12-10-24.jpg
3. In the window Edit Camera Rig Model, next to the Rig Model Name, click Edit.

Edit camera rigs parameters

Here is a description of the different parameters that can be edited in the Edit Camera Rig Model window:

2018-04-05_10-53-28.jpg

2018-04-05_15-37-36.jpg The number of cameras in the rig is fixed by the rig model and cannot be changed.

2018-04-05_15-37-58.jpg The reference camera is the camera with respect to which the other cameras are positioned. In Pix4Dmapper, having parameters that describe how the cameras are connected is crucial for calibrating camera rigs successfully. Those parameters include the spatial position of the cameras with respect to the reference one, called the rig relatives

  • The difference in translations along the x, y, z-axes (columns Transl X, Transl Y, Transl Z [mm] in the window Edit Camera Rig Model).
  • The difference in rotation (columns Rot X, Rot Y, Rot Z [Degree] in the window Edit Camera Rig Model).

2018-04-05_15-38-10.jpg Several rig processing modes are available:

2018-04-05_12-22-38.jpg

  • Rely on relatives: To be used if the rig relatives (translations and rotations) should stay fixed and not be optimized/modified.
  • Optimize Relative Rotation: Will find matches across all cameras and optimize the rotations accordingly (the translations are fixed).
  • Optimize Relative Rotation with a subset of secondaries: Similar to Optimize Relative Rotation, but reducing the computations to decrease processing time. This is recommended for multispectral cameras in Pix4D's camera model database where the results are similar to Optimize Relative Rotation (but faster).
  • Optimize Independently: The rig constraints are relaxed during the last step of calibration, to try to help for possible vibrations. This would make sense for rigs with larger cameras.
  • Disable rig: Calibrate ignoring the rig constraints.
  • Calibrate rig: Option to obtain the rig constraints (mostly for use internal to Pix4D). Requires GCPs for scale.
Was this article helpful?
0 out of 1 found this helpful

Article feedback (for troubleshooting, post here)

0 comments

Please sign in to leave a comment.