Computation Parameter

The 3D computation algorithm can be controlled by the following parameters to achieve the desired result.

• Compute Param: This field allows users to select what camera parameters to compute. For parameters not being computed, their current values are assumed to be accurate in the computation. Refer to Tips for 3D Computing for indications of how to select a proper value.
  • (none): Photo4D does not compute any imaging parameters. It computes point coordinates using the current imaging parameters. It can be used to compute the coordinates of newly added points, which the imaging parameters have obtained in the previous computation.
  • R: Photo4D computes only the camera look-at direction (Rx, Ry, Rz) for each image. This mode is usually used in a motion capture project, in which camera locations are fixed but are rotated to follow a moving subject. In this case, the TR(F) or TRF mode can be used to compute camera locations for the first frame. For the other frames, only R needs to be computed.
  • RF: Photo4D computes the camera look-at direction and focal length (Rx, Ry, Rz, F) for each image. Similar to R mode, this mode is usually used when cameras are rotated and zoomed in or out to follow a moving object during motion capture.
  • TR: Photo4D computes the camera location and look-at direction (Tx, Ty, Tz, Rx, Ry, Rz) for each image. This is the default mode and is the recommended mode for initial computation. It is generally robust since the algorithm does not need to estimate many parameters. In addition, all the other imaging parameters can be obtained easily in most cases. For example, the focal length for each image can be read from your camera when taking the image. The default value (1.0) can be used for the distortion factor (K) and all the imaging parameters usually have their default values. Thus you are strongly recommended to write down the focal length used for each image when you take the images, especially if your camera has a zoom function.
  • TR(F): Photo4D computes camera location, look-at direction (Tx, Ty, Tz, Rx, Ry, Rz) for each image and a common focal length. Employ this mode if you used a lens with a fixed focal length.
  • TR(FK): Photo4D computes the camera location, look-at direction (Tx, Ty, Tz, Rx, Ry, Rz) for each image, and common focal length and lens distortion factor (F, K) for all the images. This mode should be used when a fix lens with large distortion, e.g., a wide angle lens, was used when taking pictures.
  • TR(A): Photo4D computes the camera location, look-at direction (Tx, Ty, Tz, Rx, Ry, Rz) for each image, and a common pixel aspect (A). The current values for all the other imaging parameters are assumed accurate. This mode is used only if you use a special camera which does not have 1:1 aspect for pixel, and you do not know the accurate value of the pixel aspect.
  • TRF: Photo4D computes the camera location, look-at direction, and focal length (Tx, Ty, Tz, Rx, Ry, Rz, F) for each image. The current values for all the other imaging parameters are assume accurate. This mode needs to be used if you do not know the focal length for each images, and if a different focal length is used for each image.
  • TRFK: Photo4D computes the camera location, look-at direction, focal length, and lens distortion (Tx, Ty, Tz, Rx, Ry, Rz, F, K) for each image independently. The current values for the other imaging parameters are assume accurate. This mode can be used if you forgot to write down the focal length when you were taking the images. In addition, you can use it if you find that a straight line on the object is warped in images, which suggests a large lens distortion.
  • TRA: Photo4D computes the camera location, look-at direction and pixel aspect (Tx, Ty, Tz, Rx, Ry, Rz, A) for each image. The current values for all the other imaging parameters are assumed accurate. This mode is used only if you use more than one special cameras which does not have 1:1 aspect for pixel, and you do not know the accurate values of the pixel aspects.
• Max Iterations: the maximum number of iterations permitted in 3D computation process.
• Min Error: the minimum error reduction before Photo4D stops computing. Notice that a number "x" in the field actually means an error of "1.0E-x".
• Initial Computation: It tells Photo4D if it should add an additional pre-processing step to adjust initial imaging parameters. It should be turned on when the 3D computation is invoked the first time. It may be turned off to speed up computation for the subsequent computation. However, it may be left on all the time since it only slows down computation slightly.
• Use Constraints: It tells Photo4D if the geometric constraints should be used in computation.
• Computed P as Ref: The computation will use, as the reference points, all the points whose x,y,z values have been computed in the previous computation.
• Calibration points: provides the option to either use all points or just selected points (by Calib Point subpage in the SetUp page) to calibrate the camera.

These parameters should be setup in the 3D Computation Dialog before 3D computation.