When selecting the value for "Compute Param" field, user should select a mode that computes as fewer number of parameters as possible. For example, if the focal lengths are known for all images, then "TR" should be used instead of TRF. In "TR" mode, Photo4D only needs to compute six unknown values for each image while it has to compute seven in "TRF" mode. If the focal lengths are unknown, but all the images were taken with a similar focal length, then TR(F) should be used instead of "TRF".
More parameters need more reference information to compute. Without sufficient reference points, selecting more parameters to compute may lead to a bad result.
It is recommended to repeat the 3D computation a few times until the errors shown in the image list do not decrease any further. This is particularly helpful when the real values for an image parameter is outside its initial setup range. In this case, repeating computation can gradually adjust the parameter to its real value.
A 3D computation may be unsuccessful, i.e., the camera parameters and point coordinates computed by Photo4D are not correct. An unsuccessful computation can be caused by problems in the following areas:
One indication of a successful computation is small error relative to the size of the object for each image shown in the "Image Info" listbox. If the errors are large, say (>0.1% of the object size), it is an indication of possible problems in marking, reference point setup, and camera initial values.
It is always recommended to examine the computed locations of the cameras by opening the Camera Parameter Window. If the computed locations and look-at directions seem reasonable, it is a good indication of a successful 3D computation.
It is recommended to examine the errors of the computed point coordinates (see Check Point Information).
Yet another way to examine the computation result is to see the computed 3D locations of points in the Preview Window.