Settings

Choose aerotriangulation estimation methods and advanced settings.

Aerotriangulation settings

Presets

Presets manage all settings accessible in the user interface and may include low level settings used to control accurately aerotriangulation algorithms.

Presets can be used to save and load user settings or to setup predefined configurations.

Activating the alternate aerotriangulation engine

Default engine

Use this preset to activate the latest ContextCapture engine.

Legacy engine

Use this preset to activate the engine of previous updates.

Complete incremental engine

Use this preset to allow the engine to start from the current result, if any, to save missing photos or connect some existing components.

Keypoints density

Keypoints density value can be changed to manage a specific dataset:

Normal: advised for most datasets.
High: increase the number or keypoints, advised to match more photos in cases where the subject has insufficient texture or the photos are small. This setting makes the aerotriangulation processing slower.

We recommend you to try the Normal mode first.

Target extraction

Defines the type of targets to be extracted from the images.

Disabled: no target will be extracted.
QR Codes: Regular QR codes and ContextCapture User Tie Point QR codes will be used to create user tie points. ContextCapture Control Point QR codes will be used to create control points.
April Tags: will be used to create user tie points or to mark an existing control point.
Chili Tags: will be used to create user tie points or to mark an existing control point.

Pair selection mode

Tie points pairs may be computed using different selection algorithms:

Default: selection is based on several criteria, among which similarity between images.
Similar images only: estimate relevant pairs by keypoint similarity. It gives good results with a reasonable computation time when image similarity is discriminative enough.
Exhaustive: use all possible pairs, advised in cases where the overlap between photos is limited (e.g., for a camera rig). Exhaustive selection computation is a lot more intensive (quadratic rather than linear), so it should be reserved to a small number of photos (a few hundreds).
Sequence: use only neighbor pairs within the given distance, advised for processing a single sequence of photos if the Default mode has failed. The photo insertion order must correspond to the sequence order.
Loop: use only neighbor pairs within the given distance in a loop, advised for processing a single loop of photos if the Default mode has failed. The photo insertion order must correspond to the sequence order.

We recommend you to try the Default mode first.

Component construction mode

The component construction algorithm can be changed to manage specific datasets:

One-pass: advised for most datasets.
Multi-pass: advised only if the One-pass mode fails to include a large proportion of photos in the main component. The Multi-pass mode takes significantly more computation time.

We recommend to try the One-pass mode first. This mode is only available when using the legacy engine.

Blockwise color equalization

If enabled, an automatic color equalization will be computed for all input photographs during the aerotriangulation step. This equalization calculates the best way of ensuring consistency of appearance between all photographs, even in cases with great radiometric differences between images.

Note that this is a two-step process, needing both:

Activating the blockwise color equalization during the aerotriangulation.
Choosing "Blockwise" as the color equalization mode in the Reconstruction step.

For more details and an example result, please see the Reconstruction Processing Settings.

Splats

If enabled, splats will be processed during aerotriangulation. Splats enrich the block representation in the 3D views. See also 3D view.

Estimation methods

Estimation policy can be selected for the different block properties according to available data in the block.

Possible estimation behaviors for the different properties involved in the aerotriangulation are:

Compute: estimate without using any input estimate,
Adjust: estimate by adjusting the input estimate (according to the adjusted property, additional options can be proposed to manage the degree of freedom),
Adjust within tolerance: estimate by adjusting the input estimate while staying close to the input estimate up to a user-defined tolerance,
Keep: use the input estimate as is.

Optical properties estimation mode

We recommend to try the One-pass mode first. The Multi-pass mode is useful when AT fails because the initial parameters are far from real values (e.g. unknown focal length or unknown large distortion). It takes significantly more computation time. If possible, prefer the faster and more robust method consisting in importing photogroup optical properties estimated with some reference dataset.

Estimation groups

Set this option to ignore the photogroup structure and estimate camera properties for each photo.

May be required for image acquisition with varying zoom/focal length, or with a multi-camera system (camera rig).

This option affects aerotriangulation accuracy if used unnecessarily.

Using this option will result in an output block with one photogroup per photo.

Rig Synchro

If a rig has been defined, an additional "Rig Synchro" entry is displayed. Use "Strict" if the synchronization between the different cameras of the rig is perfect, otherwise use "Loose".

Low-level settings

Low-level settings can be set only through loaded presets. They can control directly all aerotriangulation processing settings.

Aerotriangulation presets may be provided by the technical support team to solve specific issues.