20. Working with Point Clouds

20.1. Introduction to Point Clouds

What is A Point Cloud?

A point cloud is a three-dimensional image of a space made up of many individual of data points (up to billions, even trillions). Each of the points has an x, y and z coordinate. Depending on the capture method, point clouds usually also have additional attributes that came from the capture, such as color values or intensity. These attributes can be used, for instance, to display point clouds in different colors. In QGIS, a point cloud can be used to generate a three-dimensional image of the landscape (or of another space).

Supported Formats

QGIS supports the data formats Entwine Point Tile (EPT) and LAS/LAZ. To work with point clouds, QGIS always saves the data in EPT. EPT is a storage format that consists of several files stored in a common folder. To allow quick access to the data, EPT uses indexing. For more information on the EPT format, see entwine homepage

If the data is in LAS or LAZ format, QGIS will convert it to EPT when it is loaded for the first time. Depending on the size of the file, this may take some time. In this process, a subfolder is created in the folder in which the LAS/LAZ file is located according to the scheme ept_ + name_LAS/LAZ_file. If such a subfolder already exists, QGIS loads the EPT immediately (which leads to a reduced loading time).

Worth Knowing

In QGIS it is not (yet) possible to edit point clouds. If you want to manipulate your point cloud, you can use CloudCompare, an open source point cloud processing tool. Also the Point Data Abstraction Library (PDAL - similar to GDAL) offers you options to edit point clouds (PDAL is command line only).

Due to the large number of data points, it is not possible to display an attribute table of point clouds in QGIS. However, the identify Identify tool supports point clouds, so you can display all attributes, even of a single data point.

If you want to create a new layer, in the same or another supported format, from an existing point cloud layer, see Creating new layers from an existing layer.

20.2. Point Clouds Properties

The Layer Properties dialog for a point cloud layer offers general settings for the layer and its rendering. It also provides information about the layer.

To access the Layer Properties dialog:

  • In the Layers panel, double-click the layer or right-click and select Properties… from the context menu;

  • Go to Layer ► Layer Properties… menu when the layer is selected.

The point cloud Layer Properties dialog provides the following sections:

metadata Information

system Source

symbology Symbology[1]

3d 3D View[1]

rendering Rendering

elevationscale Elevation[1]

editMetadata Metadata

basicStatistics Statistics

[1] Also available in the Layer styling panel

Note

Most of the properties of a point cloud layer can be saved to or loaded from a .qml file using the Style menu at the bottom of the properties dialog. More details at Save and Share Layer Properties

20.2.1. Information Properties

The metadata Information tab is read-only and represents an interesting place to quickly grab summarized information and metadata on the current layer. Provided information are:

  • General such as name in the project, source path, last save time and size, the used provider

  • custom properties, used to store in the active project additional information about the layer. Default custom properties may include layer notes. More properties can be created and managed using PyQGIS, specifically through the setCustomProperty() method.

  • Based on the provider of the layer: extent and number of points

  • The Coordinate Reference System: name, units, method, accuracy, reference (i.e. whether it’s static or dynamic)

  • Metadata delivered by the provider: creation date, version, data format, scale X/Y/Z, …

  • Picked from the editMetadata Metadata tab (where they can be edited): access, extents, links, contacts, history…

../../../_images/point_cloud_information.png

Fig. 20.1 Point cloud information tab

20.2.2. Source Properties

In the system Source tab you can see and edit basic information about the point cloud layer:

../../../_images/point_cloud_source.png

Fig. 20.2 Point cloud source tab

  • Settings: Set a Layer name different from the layer filename that will be used to identify the layer in the project (in the Layers Panel, with expressions, in print layout legend,…)

  • Assigned Coordinate Reference System (CRS): Here you can change the layer’s assigned Coordinate Reference System, selecting a recently used one in the drop-down list or clicking on setProjection set Projection Select CRS button (see Coordinate Reference System Selector). Use this process only if the CRS applied to the layer is a wrong one or if none was applied.

  • Provider Feature Filter: allows to restrict data to use in the current project to a subset, using functions and expression at the PDAL data provider level. Press the Query Builder button at the bottom to start setting the filter.

    ../../../_images/point_cloud_querybuilder.png

    Fig. 20.3 Filtering point cloud layer to buildings at specific altitude

    In the bottom part of the dialog, you can type a Provider specific filter expression. Such an expression can be built thanks to:

    • Fields: the list contains all the attributes of the point cloud layer. To add an attribute to the expression, double-click its name or just type it into the text box.

    • Values: frame lists values or statistics of the currently selected attribute, as a key:value pair. To add a value to the expression field, double click its name in the list: depending on the type of attribute, either the key or the value will be added to the expression. Or type the value in the expression text box.

    • Operators: this toolbar contains all usable operators. To add an operator to the expression field, click the appropriate button. Relational operators ( = , > , …) and logical operators (AND, OR, …) are available.

    The Test button helps you check the syntax of your query. Use the Clear button to wipe the query and revert the layer to its original state (i.e., fully load all the points in the layer). It is possible to Save… the query as a .QQF file, or Load… the query from a file into the dialog.

    When a filter is applied, QGIS treats the resulting subset as if it were the entire layer. For example if you applied the filter above for filtering buildings, you can not e.g. display, query, save or edit points that are of vegetation classification because they are not part of the subset.

    Tip

    Filtered layers are indicated in the Layers Panel

    In the Layers panel, a filtered layer is listed with a indicatorFilter Filter icon next to it indicating the query used when the mouse hovers over the icon. Double-click the icon to open the Query Builder dialog for edit. This can also be achieved through the Layer ► Filter… menu.

20.2.3. Symbology Properties

In the symbology Symbology tab the settings for the rendering of the point cloud are made. In the upper part, the settings of the different feature renderers can be found. In the lower part, there are sections with which general settings for the entire layer can be made and which apply over feature renderers.

20.2.3.1. Feature Rendering types

There are different options for rendering point clouds that can be selected using the drop-down menu at the top of the Symbology tab (see Fig. 20.4):

  • pointCloudExtent Extent Only: Only a bounding box of the extent of the data is displayed; convenient for overviewing the data extent. As usual, the Symbol widget helps you configure any properties (color, stroke, opacity, sub-layers, …) you’d like for the box.

  • singlebandPseudocolor Attribute by Ramp: The data is drawn over a color gradient. See Attribute by Ramp Renderer

  • multibandColor RGB: Draw the data using red, green and blue color values. See RGB Renderer

  • paletted Classification: The data is drawn using different colors for different classes. See Classification Renderer

When a point cloud is loaded, QGIS follows a logic to select the best renderer:

  • if the dataset contains color information (red, green, blue attributes), the RGB renderer will be used

  • else if the dataset contains a Classification attribute, the classified renderer will be used

  • else it will fall back to rendering based on Z attribute

If you do not know the attributes of the point cloud, the basicStatistics Statistics tab provides a good overview of which attributes are contained in the point cloud and in which ranges the values are located.

For each renderer, you can improve the data display adjusting the point symbol size or enabling surface triangulation.

../../../_images/point_cloud_symbology_overview.png

Fig. 20.4 Point cloud symbology tab

Attribute by Ramp Renderer

With singlebandPseudocolor Attribute by Ramp, the data can be displayed by numerical values over a color gradient. Such numerical values can be, for example, an existing intensity attribute or the Z-value. Depending on a minimum and a maximum value, the other values are spread to the color gradient via interpolation. The distinct values and their assignment to a certain color are called “color map” and are shown in the table. There are various setting options, which are described below the figure.

../../../_images/point_cloud_attribute_by_ramp.png

Fig. 20.5 Point cloud symbology tab: Attribute by Ramp

  • Min and Max define the range that is applied to the color ramp: the Min value represents the left, the Max value the right end of the color ramp, the values in between are interpolated. By default QGIS detects the minimum and the maximum from the selected attribute but they can be modified. Once you have changed the values, you can restore the defaults by clicking on the Load button.

  • The Interpolation entry defines how values are assigned their color:

    • Discrete (a <= symbol appears in the header of the Value column): The color is taken from the closest color map entry with equal or higher value

    • Linear The color is linearly interpolated from the color map entries above and below the pixel value, meaning that to each dataset value corresponds a unique color

    • Exact (a = symbol appears in the header of the Value column): Only pixels with value equal to a color map entry are applied a color; others are not rendered.

  • The Color ramp widget helps you select the color ramp to assign to the dataset. As usual with this widget, you can create a new one and edit or save the currently selected one.

  • The Label unit suffix adds a label after the value in the legend, and the Label precision controls the number of decimals to display.

The classification Mode helps you define how values are distributed across the classes:

  • Continuous: Classes number and color are fetched from the color ramp stops; limits values are set following stops distribution in the color ramp (you can find more information on stops in Setting a Color Ramp).

  • Equal interval: The number of classes is set by the Classes field at the end of the line; limits values are defined so that the classes all have the same magnitude.

The classes are determined automatically and shown in the color map table. But you can also edit these classes manually:

  • Double clicking in a Value in the table lets you modify the class value

  • Double clicking in the Color column opens the Color Selector widget, where you can select a color to apply for that value

  • Double clicking in the Label column to modify the label of the class

  • Right-clicking over selected rows in the color table shows a contextual menu to Change Color… and Change Opacity… for the selection

Below the table there are the options to restore the default classes with Classify or to manually symbologyAdd Add values or symbologyRemove Delete selected values from the table.

Since a customized color map can be very complex, there is also the option to fileOpen Load an existing color map or to fileSaveAs Save it for use in other layers (as a txt file).

If you have selected Linear for Interpolation, you can also configure:

  • checkbox Clip out of range values By default, the linear method assigns the first class (respectively the last class) color to values in the dataset that are lower than the set Min (respectively greater than the set Max) value. Check this setting if you do not want to render those values.

  • Legend settings, for display in the Layers panel and in the layout legend. Customization works the same way as with a raster layer (find more details at Customize raster legend).

RGB Renderer
../../../_images/point_cloud_rgb.png

Fig. 20.6 The point cloud RGB renderer

With the multibandColor RGB renderer, three selected attributes from the point cloud will be used as the red, green and blue component. If the attributes are named accordingly, QGIS selects them automatically and fetches Min and Max values for each band and scales the coloring accordingly. However, it is also possible to modify the values manually.

A Contrast enhancement method can be applied to the values: No Enhancement, Stretch to MinMax, Stretch and Clip to MinMax and Clip to MinMax

Note

The Contrast enhancement tool is still under development. If you have problems with it, you should use the default setting Stretch to MinMax.

../../../_images/RGB_render_surface_example.png

Fig. 20.7 Example of RGB renderer (left) combined with surface triangulation option (right)

Classification Renderer

In the paletted Classification rendering, the point cloud is shown differentiated by color on the basis of an attribute. Any type of attribute can be used (numeric, string, …). Point cloud data often includes a field called Classification. This usually contains data determined automatically by post-processing, e.g. about vegetation. With Attribute you can select the field from the attribute table that will be used for the classification. By default, QGIS uses the definitions of the LAS specification (see table ‘ASPRS Standard Point Classes’ in the PDF on ASPRS home page). However, the data may deviate from this schema; in case of doubt, you have to ask the person or institution from which you received the data for the definitions.

../../../_images/point_cloud_classification.png

Fig. 20.8 The point cloud classification renderer

In the table all used values are displayed with the corresponding color and legend. At the beginning of each row there is a checkbox check box; if it is unchecked, this value is no longer shown on the map. With double click in the table, you can modify:

  • the Color: the Color Selector widget opens

  • the Size: assigning a size of 0 to a category will revert it to use the default point size set for the layer

  • the Value

  • the Legend

  • the Percentage column will show you the representation of the category within the layer.

You can also right-click one or more rows to open a context menu with the options to change the color, opacity or point size.

Below the table there are buttons with which you can change the default classes generated by QGIS:

  • With the Classify button the data can be classified automatically: all values that occur in the attributes and are not yet present in the table are added

  • With symbologyAdd Add and symbologyRemove Delete, values can be added or removed manually

  • Delete All removes all values from the table

Hint

In the Layers panel, you can right-click over a class leaf entry of a layer to quickly configure visibility of the corresponding features.

../../../_images/clasification_render_surface_example.png

Fig. 20.9 Example of classification renderer (left) combined with surface triangulation option (right)

20.2.3.2. Point Symbol

Under Point Symbol, the size and the unit (e.g. millimeters, pixels, inches) with which each data point is displayed can be set. Either Circle or Square can be selected as the style for the points.

20.2.3.3. Render as a surface (Triangulate)

Check checkbox Render as surface (Triangulate) to enable the triangulation of the point cloud layer in the 2D view. This option allows rendering triangles instead of points. Each point keeps its color for interpolation in the triangle. You can control the horizontal length of computed triangles:

By checking the checkbox Skip triangles longer than option and setting up the threshold value, you can control the maximum lenght of a side of the triangles to consider in the horizontal plan. This can be particularly useful if you want to identify actual holes in the data.

../../../_images/shading_render_surface_example.png

Fig. 20.10 Rendering data as a surface with map shading (left) and with map shading, filtering large triangles (right)

20.2.3.4. Layer Rendering

In the Layer Rendering section you have the following options to modify the rendering of the layer:

  • Draw order: allows to control whether point clouds rendering order on 2d map canvas should rely on their Z value. It is possible to render :

    • with the Default order in which the points are stored in the layer,

    • from Bottom to top (points with larger Z values cover lower points giving the looks of a true ortho photo),

    • or from Top to bottom where the scene appears as viewed from below.

  • Maximum error: Point clouds usually contains more points than are needed for the display. By this option you set how dense or sparse the display of the point cloud will be (this can also be understood as ‘maximum allowed gap between points’). If you set a large number (e.g. 5 mm), there will be visible gaps between points. Low value (e.g. 0.1 mm) could force rendering of unnecessary amount of points, making rendering slower (different units can be selected).

  • Opacity: You can make the underlying layer in the map canvas visible with this tool. Use the slider to adapt the visibility of your layer to your needs. You can also make a precise definition of the percentage of visibility in the menu beside the slider.

  • Blending mode: You can achieve special rendering effects with this tool. The pixels of your overlaying and underlying layers are mixed through the settings described in Blending Modes.

  • Eye dome lighting: this applies shading effects to the map canvas for a better depth rendering. Rendering quality depends on the draw order property; the Default draw order may give sub-optimal results. Following parameters can be controlled:

    • Strength: increases the contrast, allowing for better depth perception

    • Distance: represents the distance of the used pixels off the center pixel and has the effect of making edges thicker.

20.2.4. 3D View Properties

In the 3d 3D View tab you can make the settings for the rendering of the point cloud in 3D maps.

20.2.4.1. 3D Rendering modes

Following options can be selected from the drop down menu at the top of the tab:

  • No Rendering: Data are not displayed

  • Follow 2D Symbology: Syncs features rendering in 3D with symbology assigned in 2D

  • singleColor Single Color: All points are displayed in the same color regardless of attributes

  • singlebandPseudocolor Attribute by Ramp: Interpolates a given attribute over a color ramp and assigns to features their matching color. See Attribute by Ramp Renderer.

  • multibandColor RGB: Use different attributes of the features to set the Red, Green and Blue color components to assign to them. See RGB Renderer.

  • paletted Classification: differentiates points by color on the basis of an attribute. See Classification Renderer.

../../../_images/point_cloud_3d_view.png

Fig. 20.11 The point cloud 3D view tab with the classification renderer

20.2.4.2. 3D Point Symbol

In the lower part of the 3d 3D View tab you can find the Point Symbol section. Here you can make general settings for the entire layer which are the same for all renderers. There are the following options:

  • Point size: The size (in pixels) with which each data point is displayed can be set

  • Maximum screen space error: By this option you set how dense or sparse the display of the point cloud will be (in pixels). If you set a large number (e.g. 10), there will be visible gaps between points; low value (e.g. 0) could force rendering of unnecessary amount of points, making rendering slower (you can find more details at Symbology Maximum error).

  • Point budget: To avoid long rendering, you can set the maximum number of points that will be rendered

  • Check checkbox Render as surface (Triangulate) to render the point cloud layer in the 3D view with a solid surface obtained by triangulation. You can control dimensions of the computed triangles:

    • checkbox Skip triangles longer than a threshold value: sets in the horizontal plan, the maximum length of a side of the triangles to consider

    • checkbox Skip triangles taller than a threshold value: sets in the vertical plan, the maximum height of a side of the triangles to consider

  • checkbox Show bounding boxes: Especially useful for debugging, shows bounding boxes of nodes in hierarchy

20.2.5. Rendering Properties

Under the Scale dependent visibility group box, you can set the Maximum (inclusive) and Minimum (exclusive) scale, defining a range of scale in which features will be visible. Out of this range, they are hidden. The mapIdentification Set to current canvas scale button helps you use the current map canvas scale as boundary of the range visibility. See Visibility Scale Selector for more information.

Note

You can also activate scale dependent visibility on a layer from within the Layers panel: right-click on the layer and in the contextual menu, select Set Layer Scale Visibility.

../../../_images/point_cloud_rendering.png

Fig. 20.12 The point cloud rendering tab

20.2.6. Elevation Properties

In the elevationscale Elevation tab, you can set corrections for the Z-values of the data. This may be necessary to adjust the elevation of the data in 3D maps and its appearance in the profile tool charts. There are following options:

  • Under Elevation group:

    • You can set a Scale: If 10 is entered here, a point that has a value Z = 5 is displayed at a height of 50.

    • An offset to the z-level can be entered. This is useful to match different data sources in its height to each other. By default, the lowest z-value contained in the data is used as this value. This value can also be restored with the refresh Refresh button at the end of the line.

  • Under Profile Chart Accuracy, the Maximum error helps you control how dense or sparse the points will be rendered in the elevation profile. Larger values result in a faster generation with less points included.

  • Under Profile Chart Appearance, you can control the point display:

    • Point size: the size to render the points with, in supported units (millimeters, map units, pixels, …)

    • Style: whether to render the points as Circle or Square

    • Apply a single Color to all the points visible in the profile view

    • Check checkbox Respect layer’s coloring to instead show the points with the color assigned via their 2D symbology

    • unchecked Apply opacity by distance from curve effect, reducing the opacity of points which are further from the profile curve

../../../_images/point_cloud_elevation.png

Fig. 20.13 The point cloud elevation tab

20.2.7. Metadata Properties

The editMetadata Metadata tab provides you with options to create and edit a metadata report on your layer. See Metadata for more information.

20.2.8. Statistics Properties

In the basicStatistics Statistics tab you can get an overview of the attributes of your point cloud and their distribution.

At the top you will find the section Attribute Statistics. Here all attributes contained in the point cloud are listed, as well as some of their statistical values: Minimum, Maximum, Mean, Standard Deviation

If there is an attribute Classification, then there is another table in the lower section. Here all values contained in the attribute are listed, as well as their absolute Count and relative % abundance.

../../../_images/point_cloud_statistics.png

Fig. 20.14 The point cloud statistics tab

20.3. Virtual point cloud

Lidar surveys of larger areas are often multi-terabyte datasets with many billions of points. Having such large datasets represented as a single point cloud file is not practical due to the difficulties of storage, transfer, display and analysis. Point cloud data are therefore typically stored and distributed split into square tiles (e.g. 1km x 1km), each tile having a more manageable file size (e.g. ~200 MB when compressed).

Tiling of data solves the problems with size of data, but it introduces issues when processing or viewing an area of interest that does not fit entirely into a single tile. Users need to develop workflows that take into account multiple tiles and special care needs to be taken to deal with data near edges of tiles to avoid unwanted artefacts in outputs. Similarly, when viewing point cloud data, it becomes cumbersome to load many individual files and apply the same symbology.

Here is an example of several point cloud tiles loaded in QGIS. Each tile is styled based on min/max Z values of the tile, creating visible artefacts on tile edges. The styling has to be adjusted for each layer separately:

../../../_images/point_cloud_individual_tiles.png

Fig. 20.15 Individual point cloud tiles loaded, with artefacts on edges

In the GIS world, many users are familiar with the concept of virtual rasters. A virtual raster is a file that simply references other raster files with actual data. In this way, GIS software then treats the whole dataset comprising many files as a single raster layer, making the display and analysis of all the rasters listed in the virtual file much easier.

Borrowing the concept of virtual rasters from GDAL, virtual point cloud (VPC) is a file format that references other point cloud files. Software supporting virtual point clouds handles the whole tiled dataset as a single data source.

../../../_images/point_cloud_vpc.png

Fig. 20.16 The virtual point cloud

Displaying and manipulating virtual point cloud is much more fluent and easy.

../../../_images/point_cloud_vpc_2d.gif

Fig. 20.17 The virtual point cloud output on 2D: displaying details when zooming in

At the core, a virtual point cloud file is a simple JSON file with .vpc extension, containing references to actual data files (e.g. .LAS, .LAZ or .COPC files) and additional metadata extracted from the files. Even though it is possible to write VPC files by hand, it is strongly recommended to create them using an automated tool, such as:

For more details, please refer to the VPC specification that also contains best practices and optional extensions (such as overviews).