Clouds

Selecting one of the manufacturers will open the standard Carlson file selection dialog and after the file has been selected the file is imported. If an ASCII format is selected the Import Clouddialog will be displayed.

Here, the user must define what delimiter was used for the ASCII file, if the data not organized properly in columns (all the data is in one column for example); the delimiter will probably need to be changed. Clicking the header of each column will bring up a menu with all the possible values that can be assigned to it, click each column's header to set its proper value. At the top of the dialog is the Presets Panel, which allows you to save the current column assignments to be used in the future. The import column is used to determine whether to import that particular column (often files will have text or header information that isn't actually part of the data set). Save the current configuration by clicking the + button in the panel, presets can also be removed by clicking - button. Presets are saved in the global program settings so they can be used across multiple projects.

Creating a Cloud

There are several methods for creating a cloud, but all of these methods use one of two cloud creation dialogs. The most common and direct method of creating a cloud is to right-click the source data object (either another cloud, a mesh or a scan) and select Create ⇒ Cloud. This dialog operates in two different modes: multiple source mode and single source mode.

Single Source mode is typically the result of right-clicking a data object in the Point Cloud tree structure and selecting Create Cloud, this mode only allows the cloud to have the object that was selected as its source. Multiple Source mode is only accessible by right-clicking the Clouds folder itself and selecting Add ⇒ Newin the menu. In this mode, the user will see a tree structure that represents the project on the left side of the dialog. You can toggle inclusion of any object in the project by clicking the x icons next to them to turn them green. The right side of the dialog works the same as in single source mode.

Cleaning a Cloud

Right-click a cloud and select Cleanto open the Clean Cloud dialog. If you are viewing a cloud in a scene, select the part of the cloud to clean using the options in the Selection Setpanel of the Action Tab, then select Cleanfrom the Editpanel. The Clean Cloudcommand launched when viewing a scene will only use the Redundant Pointsmethod.

The Cleanfunction will attempt to clean up the data in a cloud through three different methods, Duplicate Points and Isolate Points may be used individually or in combination. To use Redundant Points the first two options must be disabled.

Remove Duplicated Pointsmethod will search for any places where two or more points in the cloud are within the Distance thresholdof each other. If any such points are found, they are deleted from the cloud. This method is to help remove redundant data that, for all intents and purposes, are the same given the current data set (say a data set that spans several miles having two points within inches of each other) and reduce the size of the data set. This can be carried out on a 2D or 3D basis.
Remove Isolated Pointsmethod will search for any points in the cloud that have less than Minimum Neighbors Countpoints within Distance Thresholdunits from them and delete them. This will remove points that are likely to be outliers, which could be a result of bad data (such as the scanner hitting a dust particle several meters off the ground). After clicking the green check button, the cloud cleaning process will begin. The time required to clean a dataset varies depending on the size of the dataset.
Redundant Pointsmethod uses a series of user parameters to clean points from the cloud. Using a box of user defined size, points are checked for fitting the plane and/or color match. If no points match the box is subdivided and the check repeated. These sub-boxes will again be subdivided if no points match. If points are found that meet the criteria a point closest to the average of those in the box is copied to the clean cloud. If no match is found, all the points from that box are copied into the clean cloud unless Strong Filtration is on.

• Box Length and WidthSets the horizontal distance to search for points to compare.
• Box heightSets the vertical component completing a 3D cube search area for points to compare.
• Residual toleranceis the measure of how far from a best fit plane points can be. A typical value would be 1/10 of the smallest box dimension.
• Color variancedetermines how much the color of a point can vary. When set to 255 color variance is not used.
• Process Planesallows the user the ability to choose to clean vertical or horizontal data only or to clean both.
  
• Recursion depthis the number of times we initial box is subdivided to find points that meet the parameters. Typical values used would be 3 or 4.
• Strong Filtrationwhen on copies a point closest to the average for plane and/or color even if no plane or color match was established.

Resampling a Cloud

Point Cloud has two methods of cloud reduction. There is a fast, naive method, and a slower, more intelligent method. Right-clicking a cloud and selecting Resamplewill bring up the Resample Clouddialog.



Stepmethod is much faster at the cost of being less intelligent. It will simply reduce the cloud to 1 / Step its current size by only keeping one out of every nth vertices. So in the case of a step size of four, it will traverse the cloud and only keep every fourth vertex, deleting three for each one it keeps.
OC-Treetree method is much more intelligent, but it can have significantly longer run times than the step method. It divides the bounding box of the cloud into blocks defined by the resolution size and then filters out points based on the minimum and maximum parameters. If a block has less than the minimum number of points, its contents will be deleted, if a block has more than maximum points, random points within that block will be removed until it has maximum points inside it.

Transforming a Cloud

Any created cloud can be easily transformed by right-clickingon the cloud and select Transform. This command will create a new cloud.


Multiple transformations can be carried out and the sequence of this can be assigned. To add a transformation, click on the 'plus' image. This will open up the following dialog:


The cloud can be:


1. Translated - shifted (positively or negatively) along the X, Y and/or Z axis
   
2. Rotated - about the X, Y and/or Z axis or CCW (counter clockwise)
3. Scaled - enter in a scale factor to be applied to the X, Y and/or Z axis
4. Advanced - enter in a the values manually

Once the values have been entered in, the sequence can be altered by selecting a type and moving the value up and down using the up and down arrows. The values can be amended by selecting a type and then clicking on the 'i'. To delete a transformation option, left-clickon the type and click on the red X. To accept the transformation, click on the green tick located at the bottom-left of the dialog box - this will create a new cloud with the additional text "Transformed" in the project manager. To cancel the operation, click on the orange X.


Changing the Projection
To create a new cloud on a different projection, right-clickon the nominated cloud and select Change Projection.


Enter in a name for the new cloud then set the Source Coordinates and units and then set the destination coordinates and click OK.


Adjust Color
By right-clickingon a cloud and selecting Adjust Colorwill open the following dialog:


Enter in a name for the new cloud. By clicking on Recolor by Image, the software will ask the user to select a Geo-referenced image to use to color the pointcloud. If the image is does not have an associated georeferenced world file, the software will use a best guess method to drape the image over the points - essential, it will amend existing RGB values or assign an RGB value to each point.

Under the Modificationsarea, the user can select Addor Multiply. By selecting Add, Red, Green, Blue and/or the Intensity values can be altered by adding a nominated value to each entity. By choosing the Multiply option, the existing values will be magnified by the values entered.


Register Clouds
Using the Registercommand will look at position differences between the Control Points (presumed to be correct in World Space) and either the Coordinate Points , or a list of Target Points from a Scan Position (either of which can be written to via picking points in Create Point mode). There are four registration methods are available; Use existing references, Match point names, Match point descriptions and Minimize point position error. If you select Minimize point position error, the Registration Settings panel will activate and you can set Maximum Error, minimum References and toggle Update target point names and Update target point descriptions.


The software will then transform the whole Cloud to align the points selected, creating a new point cloud.


Merging Clouds


Two existing clouds can be merged to create a single cloud. Right-click on one of the desired clouds and select Merge.


The user can then name the new cloud, select what attributes they would like to merge and select the clouds to be merged. More than two clouds can be merged by clicking on Add Another Cloud. Select the green tick when finished.



Tab Location(s): Project
Tree Folder: Clouds
Prerequisite: Existing Cloud, Mesh, Scan data or ASCII file

-->