GeoTalks Express: Lidar Tips and Tricks – Questions And Answers
In the latest GeoTalks Express webinar, members of the Blue Marble team reviewed and demonstrated some helpful tips and tricks to keep in mind when working with lidar data in Global Mapper Pro. If you are interested in learning more about working with point cloud data in Global Mapper Pro, consider attending one of our Lidar Data training classes.
All registered attendees of this recent webinar should be emailed a link to view the recorded session on YouTube. If you did not register for this live webinar, sign up here to see recordings of this and past webinars, and check out our YouTube channel for many more helpful videos.
Below are the questions asked to our presenters during the live webinar, with answers provided by a member of our technical support team.
When would you use the ground classification tool versus the point segmentation tool to classify ground points?
The Automatic Ground Classification tool identifies and classifies points. The Segmentation tool identifies clusters of similar points based on the input parameters but does not automatically apply a classification. If the Automatic Ground Classification tool is not working well with a particular dataset and the ground representing points have similar characteristics throughout the point cloud, the segmentation tool and manual segment classification can be used to classify ground points. The Segmentation tool can also be useful when looking to identify different ground surface or ground cover types in a point cloud.
Do you have any tutorials on the new point cloud segmentation by spectral graph partitioning?
Documentation discussing the Point Cloud Segmentation tool can be found here in the Global Mapper knowledge base. An Ask the Experts video showing the tool is available here on YouTube and a blog post outlining the use can be found on the Blue Marble blog.
Is it possible to identify water features from lidar points? I realize this is challenging because water isn’t usually captured in the returns.
Global Mapper Pro does not have a tool for automatic classification of water points, but these points can be manually selected and classified in the top-down, 3D, or Path Profile views. Often water does not construct well in lidar data and results in a rough, noisy, and inaccurate surface.
Wondering why the automatic classification of buildings didn’t pick up more of the building points? Seems that viewing them in the profile view shows them as building points. Does it make more sense to go back and rerun the auto-classification with new values before doing manual editing?
The Path Profile view can be deceiving because it makes a depth of points in a cross-section appear flat. Some of the unclassified points around the top of the classified buildings may be smaller changes in roof elevation due to vents and other structures. These points are sometimes left out because they do not adhere to the flatter roof structure Global Mapper Pro looks for in automatic classification. You can absolutely run the automatic classification again to try and achieve better results before moving on to manual editing. You should also keep in mind the goal of the workflow. If you are not looking to model all small structures on top of a building and simply want to extract building footprints, a complete classification of every point may not be necessary.
Can you plot a profile of distance vs. intensity (rather than elevation)?
A Path Profile is a cross-sectional view of the 3D data, so it always shows distance horizontally and elevation vertically. While this view looks like a graph, it cannot be used to show intensity value vertically with distance horizontally. That being said, you can apply different Lidar Draw Modes to point clouds in the Profile View to draw the point returns by intensity while viewing the cross-sectional perspective.
Was the vectorization file for both the buildings and the water generated manually? Or did they do it automatically? The quality is very good to be automatic.
Global Mapper Pro does contain an Automatic Feature Extraction tool to create vector features from a classified point cloud. This tool can be used to extract vector building features like the ones shown in the webinar. Options for simplification and regularization of the building shapes work well to clean up the extracted building vectors in order to produce appealing results.
There is no automatic option to extract water features from a point cloud, and the feature used in the webinar is from a state hydrology dataset. That being said, water representing points can be manually classified, selected, and used alone in the creation of a bounding polygon to derive a water body area feature from lidar data.
Can you get the roof angles in adding 3D data to the vector buildings? They were all flat roofs in the example.
The buildings appeared flat after applying elevations because a single elevation was added to each building area feature. To have varying elevations for the buildings per vertex, elevations would need to be added to the features instead of a single elevation. This can be done with the Apply Elevations to Selected Features digitizer option but often results in a messier-looking end product.
Alternatively, building roof planes can be extracted from a classified point cloud using the Automatic Feature Extraction tool in Global Mapper Pro.
What if I have very steep terrain, would the option “Color Lidar by Height Above Ground” still work that well?
Height above ground values used for visualization and analysis in Global Mapper Pro are calculated on the fly when the Color Lidar by Height Above Ground option is first selected for a dataset. To do this, a minimum value binning method is used under the hood to calculate a rough ground model from the data and then used to derive the height above ground for each point. This method works for both flat and hilly terrain as the ground model generated under the hood is based on the point spacing of the point cloud and the elevation of the point returns in local areas.
Are the height values always in meters in the query, or if the point cloud is in feet, then the height would be feet?
The height above ground Lidar Draw Mode always calculates values in meters regardless of the native vertical units of the data. This is similar to how point spacing and density are calculated in metric units.
When assigning height attributes to buildings, how does Global Mapper determine the Z value for the “HEIGHT” option?
With vector features, some common elevation attribute names (like Elevation, Elev, Height, and Z) are automatically recognized and used by Global Mapper to display features with a height or elevation. You can also specify an attribute to use for elevation in a layer by selecting a specific attribute from the Elevations tab of the layer options. To go along with this, an altitude mode is designated to guide Global Mapper on how to interpret the elevation values.
This was collected by a multispectral drone?
The data used in the webinar was created using the Pixels to Points tool with multispectral drone-collected images. Global Mapper Pro’s Pixels to Points tool takes a set of images and with the structure from motion process, generates a point cloud, orthoimage, and/or 3D mesh data layers.
Have you ever filtered out vegetation using the NDVI?
After calculating NDVI values for a multispectral point cloud, these values can be used like any other attribute to filter or search the returns in a point cloud. Personally, I have not often worked with multispectral data, but this is a wonderful use case for it.
Could you create some more custom shaders and make them available as options within the software?
The built-in shader options in Global Mapper allow the terrain to be shown in a number of visually appealing and clear ways. There are no current plans to expand the available shaders, but users can create custom shaders through Configuration > Shader Options and share created shaders with colleagues and other Global Mapper users by Exporting User Settings.
I missed how you copied selected items to a new layer, was that a right-click once selected?
Features can be copied and pasted to a new layer after selection with the digitizer. To copy and paste, use the keyboard shortcuts ctrl + c and ctrl + v, or access these functions from the Edit menu in Global Mapper.
When applying color from RGB imagery into the point cloud, does it do a good job at applying color to side faces of buildings or more vertical walls?
Since a raster image is not a 3D layer, colors are applied to point cloud returns from a top-down perspective. Vertical faces, like the sides of buildings, will likely inherit the same colors from the image as they have little to no horizontal extent, and building sides are likely not captured from an aerial view.
Can it be done in reverse? Extract an image from RGB lidar data?
Yes, RGB color values, and other lidar attributes can be gridded in the Create Elevation Grid tool. Changing the Grid Type in the Grid Creation Setup allows you to create an image grid based on the RGB, intensity, or other attributes present for the point cloud.