Processing Settings for SLAM Scanner

123-2026-04-28-888

SLAM uses a variety of sensors, primarily LiDARLight Detection and Ranging, to map the space around you, identify features and surfaces and then determines the scanner's spatial relationship to those features as it moves through an environment. This is also reinforced with other types of sensors like an IMUInertial Measurement Unit. In addition to Orbis and Orbis Premium captures, you can also process captures made by GeoSLAM scanners.

On this page you can define settings for SLAM scan processing. These settings are applied when:

• You import a .geoslam file to create a new project.

• You import a .geoslam file into an existing project.

• You reprocess an existing project with SLAM.

If you run the processing several times, keep in mind that the processing always starts from parameter Capture environment. In other words, you can deactivate Always run SLAM, but you may have to adapt the other parameters on this page. For example if you run a SLAM processing with a rigid transformation and specific colorization options as a first step, you do not need the SLAM processing again if you want to apply a non-rigid transformation.

SLAM parameters

Always run SLAM

If activated, SLAM processing is always done. You can deactivate this switch if you do not need SLAM processing, e.g. if you have already run a SLAM processing before.

Capture environment 

In dropdown Capture environment, select a processing type with information about the environment where the scan was made. This can aid the processing quality.

• Confined: An environment where the system will be exposed to surfaces at a maximum 1.5m from the scanners trajectory. To be used in a tight indoor environment or if temporarily moving through a tight space.

• Forest: Dense forest environment where irregular features occur on all 6 sides of the scanner (left, right, top, bottom, front, back). The start and finish position are the same.

• Indoor: Indoor environments where features are usually planar on all sides of the sensor and maximum range rarely exceeds 30m.

• Open Loop: Various environment where the start and finish positions are not the same. To be used in linear scans where capture does not exceed 20 - 25 minutes.

• Standard: The default setting for a feature-rich environment. Feature-rich environments are those where unique objects/ surfaces surround the system on 5 sides (front, back, left, right, bottom). These objects/surfaces should be at maximum 40m from the system.

• Tunnel: A feature-poor environment where the system is surrounded by irregular surfaces on 4 sides (left, right, top, bottom). These surfaces should be at maximum 25m from the system.

• UAV: Airborne missions taken by a drone where features are no more than 40m from the sensor. Bounding box has been modified to remove UAV system from the processing data. This setting removes the drone rotor blades from the capture. Currently, UAV is not possible with Orbis, you can only use ZEB Horizon without Vision.

• Vehicle: Environment where the vehicle is moving at no more than 30 mph (48 km/h). The start and finish positions should be approximately in the same place (some 2-3m). Features vary in distance from the trajectory.

Raw points used (Beta)

This option allows you to decide if you want a dense point cloud which will increase the processing time, or a less dense point cloud which will reduce the processing time. Note that Revo captures have a lower point density by default. Therefore, the point reduction of the Raw points used feature will not affect these point clouds.

All: 100% of the points are used for all scanner types. Selecting this option will increase the processing time.

Reduced: This option will process less points for all ranges. Selecting this option will reduce the processing time.

Sparse: This option will further reduce the number of processed points for all ranges. Selecting this option will further reduce the processing time.

Loop Type 

• Closed: The scan capture started and ended at the same position.

• Open: The scan capture did not start and end at the same position (linear scan).

Scanner Configuration

Mount Type

To determine the correct offset and positioning of captured static points or user events compared to the reference points, you must select the mount type which was used during scanning. For example, when using the monopod, captured static points or user events were displayed at the monopod level and not at ground level. By entering a monopod height, you make sure that the static points or user events are displayed at ground level.

• Handheld: The scanner was carried using the handheld. This is the default mount type.

• Custom: The scanner was carried using a custom mount. In this case you are asked to enter offset values for the x, y and z axes.

• Monopod: If a monopod was used during scanning, you must enter the monopod height.

Georeferencing

Transformation type

Rigid: In the rigid transformation the point cloud is transformed using a rigid translation and rotation matrix to best fit the control points. The resulting point cloud is only transformed in space and not in shape. The relative position to other points in the point cloud will remain the same. Comparing the shape of the point cloud before and after the rigid transformation should show identical results. A rigid translation is helpful as a first step to evaluate the current alignment errors.

Non-rigid: The non-rigid transformation will move every point in the point cloud so that the XYZ position of every point is altered. Additionally, the relative positions of all the points will not remain the same. Comparing the shape of the point cloud before and after the non-rigid transformation should show dissimilar results.

Control point file

File containing the known (reference) control points. This should be a text file that meets the following requirements:

Columns must be separated by commas or whitespace (spaces or tabs)

• The file must contain a header line.

• The first four columns must contain the following in order:

  • control point name

  • X

  • Y

  • Z

Control points confidence, if non-rigid 

Only necessary if you have selected transformation type Non-rigid. The control points confidence is the maximum error norm for each control point allowed by SLAM.

For example, if you set this parameter to 10mm, i.e. the maximum combined error when placing the base plate of the scanner on the control point and when measuring the control point with a total station or GNSS, etc., the SLAM processing for a non-rigid transformation will make sure that the error norm is lower than the entered control points confidence, in this case 10mm.

Extract any reflective targets

If you activate this switch, the project's reflective targets will be extracted and shown as reference points in the project structure view.

Filters

Minimum intensity, Maximum intensity

You can select the minimum and maximum limit of the intensity scale.

Minimum range (m), Maximum range (m)

Defining a range filter is used to limit the range of the data. Any data closer than Minimum range and further than Maximum range will be ignored.

Transient, Grid size

Used for moving objects. You can define the number of seconds that are not stationary which will be filtered out of the point cloud under Transient time window. Set the grid size you want under Grid size.

Threshold for number of points in a voxel

Voxels are small sections or cubes of the scan window. With this parameter, you an define the max. number of points in one voxel.

AI Outlier

Remove outlying data points from the point cloud with AI support. The filter classifies points based on local density. Compared to the former Outlier filter, the AI Outlier filter achieves much better results.

Outlier

Remove outlying data points from the point cloud. Outlying points may be caused by rain or dust or by partial reflections from edges. The Statistical Outlier routine classifies points depending on their distance from the neighboring points.

Number of neighbors

Number of neighbors referring to how many nearby points are considered when analyzing each point in the point cloud.

Standard deviation threshold

The standard deviation threshold is a measure of how much variation or spread exists in the distances between points within the point cloud. When filtering the point cloud, this parameter allows you to set a limit on how much the distances between neighboring points can vary. Points with distances exceeding this threshold are often considered outliers or noise and can be removed or marked for further processing.

If you decrease the standard deviation threshold, points with very irregular spacing will be filtered out, because they are considered as noise. If you increase the standard deviation threshold, points with a greater variation in spacing will be allowed to remain in the point cloud.

Noise Reduction

If activated, the point cloud is improved by reducing surface irregularity/noise and removing outlying data points.

Thinning filter, Grid size: If you activate this switch, this filter creates a uniform point cloud and reduces the number of points which make processor and memory significant functions run more quickly. The thinning method employed is based on a Voxel grid. Set the grid size you want under Grid size.

Flash Scans

Extract available Flash scans

If you activate this switch, available Flash scans, i.e. enhanced colored static scans, are extracted. Note that this is only possible for captures made with the Orbis and Orbis Premium scanner.

Edge sharpening

If activated, stray edge points are removed from the static scans.

Output location for Flash E57 (if required)

Click the browse button and select the directory where you want to store the Flash scans.

Minimum range (m), Maximum range (m)

See above.

Colorization

Colorize mobile data

Searches for camera images that are available in the scan and uses these images to colorize the mobile data. Depending on the mount type you have selected, the colorization mask changes for both, Flash and mobile scans.

Select mask to exclude certain areas of the image

Allows you to mask out areas of stationary color from the panoramic images (such as the scanner and the operator) which would otherwise affect the accuracy of the color projected onto the point cloud. The masking is defined by the capture technique.

• Backpack: Image mask that you can apply when you are using the backpack.

• Custom: Custom image mask that can be created for you if you don't use the backpack, the handheld mode or the monopod. Contact FARO Support (support@faro.com) for further information.

• Handheld: Image mask that you can apply when you are using the handheld mode.

• Monopod: Image mask that you can apply when you are using the monopod.

• None: No image mask will be applied.

Create Panoramic Images

If you activate this switch, panoramic images will be available after processing. Otherwise, you will only get a colorized point cloud without panoramic images. Note that the panoramic images will not be available in the quick view if you disable Show the full resolution panorama in the quick view if available in the Views settings.

The maximum distance of points from the camera to be included in coloring

You can define up to which distance from the camera points will be colorized. Uncolorized points will be deleted from the point cloud.

Image Filtering

The approximate desired distance between images

You can define the approximate distance between images shown on the trajectory. Note that only these images will be used to colorize the point cloud.