Preferences - Processing and Cameras

Process:

Field	Description
Optimization termination
Maximum iterations	Controls how many iterations optimizer will run before giving up trying to find a solution.
Solution tolerance	Controls how close the Total Error bar values have to be together before optimizer decides a solution has been reached.
Bundle Stability Measure	Some projects do not have a strong mathematical solution. Processing has trouble deciding what is the correct solution. A weak mathematical project (that is otherwise correct with correct references, good calibration, etc.) is an indication that results cannot be always trusted. In this case processing fails. The Bundle Stability Measure value itself is a matrix stability measure. It can be increased to 1e-3 from the default 1e-5 for example to make some mathematically unstable projects solve (i.e. complete processing). Long focal length calibrations are a project type that may benefit from this change.
Project Status Report
Only report parameter correlations over this %	Controls which correlation values are shown in the Project Status Report.
Default Point Standard Deviation Weights (pixels)
Marked points	Controls the one-standard-deviation (68% probability level) that a Marked Point is within its true location. The size of the final Total Error (see Final Total Error) in a processing will be affected by this value. If this value is close to the true marking precision for all points (and other things like camera parameters are good) then the final Total Error (sigma0 value) will come out close to 1.0.
Sub-pixel target points	Controls the one-standard-deviation (68% probability level) that a Sub-pixel Target Marked Point is within its true location. See the Marked points value above for the relationship to Final Total Error.
SmartPoints	Controls the one-standard-deviation (68% probability level) that a SmartPoint is within its true location. See the Marked points value above for the relationship to Final Total Error. See also SmartPoints Project.
Sphere Points	Controls the one-standard-deviation (68% probability level) that a Sphere Target Marked Point is within its true location. See the Marked points value above for the relationship to Final Total Error. See also and Sphere Target Marking.
Background Processing and Alerts
Prompt to turn off background processing if it takes longer than: (sec)	The maximum length of time to perform background processing before being prompted to switch to manual processing.
Show the % of approximated points alert	A popup alert (bottom right of screen) will appear if the check box is checked, and if there are more than the specified percentage of points calculated only by approximation.
Show could orient alert	If checked, a popup alert (bottom right of screen) will appear when enough points are referenced for possible orientation.
Show outside calibration alert	If checked, a popup alert (bottom right of screen) will appear if a point is marked outside of the calibration point coverage area (i.e. the area of the lens that was solved during calibration).
Other
New project default setting for high residual point removal optimization	This setting enables the automatic search and unreferencing of points with high residuals during processing. In addition, photographs may be set to 'do not use in processing' if they no longer meet the minimums. By default, the feature is on for Automatic Coded Target projects only. See also Total Error Dialog.
2D Template projects use Robust Optimization	The solution for 2D Template project set up uses the Smart Orient (Robust Optimization) which makes it more robust to errors in photography and targets. Some project might solve with this setting off.

Orientation:

Graphical user interface, application

Description automatically generated

Field	Description
Orientation Point Counts
Number of points to include in relative orientation	Controls the number of points chosen from all photo points to be used during relative orientation stage. A low number speeds up orientation but there is a greater chance of failure or incorrect orientation.
Number of points to include in orientation check	Controls the number of points studied to allow relative orientation pick from multiple conflicting solutions. See Valid orientation checks below.
Valid orientation checks
If roll angle is greater than __ use this weighting factor against the solution: __	These two values control how PhotoModeler picks from multiple conflicting solutions to orientation. This weights against cameras that are rolled upside down because this is uncommon.
If a projected point is closer to camera than __ use this weighting factor against the solution: __	These two values control how PhotoModeler picks from multiple conflicting solutions to orientation. This weights against cameras that placed extremely close to the model data because this is uncommon.
Perform Angle Checks	If checked, photo pair angles are checked before setting up orientation pairs so that pairs with low angles do not cause poor orientation results.
Fit Error Threshold	The threshold used to determine whether photo pairs have low angles. Generally the higher the value, the closer the ‘fit’ and therefore the lower the angle difference. Increasing the threshold causes the check to be more strict. This setting is very project dependent and may require some trial and error to find the best setting.

Camera Calibrator:

The Camera Calibrator page controls the Camera Calibration process.

Graphical user interface

Description automatically generated

Field	Description
Custom Calibration Control Positions
Four Planar Control Point Positions	The positions in 2D of the four control points that lie on a plane. Only relative size is important (that is these are not absolute values). Change this only if you are using a non-standard calibration sheet.
Auto-referencing
Required number of photos to be referenced for each point in stage one	One of the stages of calibration is an automatic referencing of points. The first time this happens the camera parameters are known only very roughly so it can be difficult to do well. To make this more robust we force each point to be referenced on at least this many photographs before being accepted. You may need to lower this if the first stage auto-referencing is failing due to too few points found.
Targets and Sub-pixel Marking
Ratio of largest target in image to image size	The size of the targets as they appear in the calibration images. This is the largest target in all images (the target closest to the camera). Ratio of image size.
Ratio of smallest target in image to image size	Like above but for the smallest target image or the size of the target farthest from the camera in all calibration photos.
Targets are black on white background	If checked the targets are black filled circles on a white background.

Include IDs when printing multi-sheet calibrations	Show the target ID when printing calibration sheets for the multi-sheet calibration process.

Self and Field Calibration:

The Process page controls those options to do with Processing and in particular Self/Full Calibration.

Graphical user interface, application

Description automatically generated

There are many settings here and their relationship to processing is complex so we give this whole section a caution sign . While many of these settings are complex they are important and for advanced projects you may find yourself adjusting some of them. If after reading the descriptions below you still require guidance please contact PhotoModeler Technologies technical support.

Field	Description
Camera Calibration (self or field)
Calibrate when greater than these number of good points on all photos	When there are more than this many 'good' points (see Preferences - Processing and Cameras) on every photo then self-calibration or full calibration can execute during optimize. See Camera Optimization.
Calibration in one camera projects is a full field calibration (otherwise self-calibration)	If checked, and there is only one Camera assigned to all photos in the project then a field camera calibration is done instead of a self-calibration. See Field Calibration.
Self-calibration uses Std. Dev. Weights	If checked then self-calibration uses the weighting factors found in the "Solve during self-calibration" section. If not checked self-calibration uses a free form un-weighted solution which may produce more consistent results but is far less stable (esp. if principal point is being solved for).

Imports/Control:

Field	Description
Use old format DXF import	Check this if you’re having difficulties with the DXF file import, as the ‘old’ system may be more compatible with the model entities.
Use alternate loading method for DXF and OBJ files	When checked, the import uses the ‘FBX’ loading system rather than PhotoModeler’s native system.
Number of decimal places in…	Controls the number of decimals shown in the list views of the Imports and Coordinate Systems Pane.
Locate photo zoom percent	The zoom percentage of the photo that opens when showing the ‘located’ point.
Locate open photos tabbed	When checked and when multiple photos show the ‘located’ point, the photos open tabbed, otherwise they’ll open tiled in the Photo Windows pane.
Show the imported ID number when displaying IDs of marked Control Points	When checked the imported ID number will be displayed along with the control point name on control points marked on photos.
Only show the imported point information when showing IDs of marked Control Points	When checked only the imported ID number will be displayed on control points marked on photos.
Default Import Material color	Set the Material color automatically assigned to imported objects.
Warn if preset scale fit is larger than (percent)	A threshold to flag potentially bad scale(s): A) Warn when a Coded Target Preset is set on a project and a scale error is larger than this value. B) During export to note scale may not be correct. C) Used by Check Distance table to bold distance errors (%) that are larger than this value.

Control Orient Inverse Camera:

Graphical user interface

Description automatically generated

Field	Description
Do exhaustive control orient search when fewer than these number of control points	Controls how many control points are used when searching for the optimal set of three control to do the initial closed form resection. The larger the number the greater chance it finds the best set but also slower.
Inverse Camera Parameter Standard Deviation Weights
Use Standard Deviation Weights during calibration of InverseCamera parameters	Single photo resection with InverseCamera does a mini-global optimization. If checked this tells that optimizer to use the standard deviation weights found below.
Inverse Camera Parameter Standard Deviation Weights	All the standard deviation values for all parameters used in Inverse Camera calibration.
Thresholds for minimum number of control points
Basic minimum	Controls how many marked control points any photo must have to be considered for control resection. This value cannot be less than three. See also Minimums.
For inverse focal length	Controls how many additional control points are required above the minimum to also solve for the camera's focal length. See also Minimums.
For inverse principal point	Controls how many additional control points are required above the minimum to also solve for the camera's principal point. Must be at least two. See also Minimums.
For inverse format size	Controls how many additional control points are required above the minimum to also solve for the camera's format size ratio. See also Minimums.
Minimum % image coverage by control points	Controls how big an area the marked control points must cover in any photo for it to be considered for control resection and InverseCamera. See also Minimums.