Author: Lewis Graham, July 8, 2021
One of the difficulties today with unmanned aerial system (UAS) LIDAR and imaging sensors is the lack of good post-processing software. It is not uncommon to be told by a vendor of a drone LIDAR system that, once the point cloud is generated by their proprietary geocoding software (usually in LAS format) you are on your own. Suggestions are made such as “buy LP360 from GeoCue, Global Mapper from Blue Marble or use some other solution”. When you encounter issues such as data too noisy for a ground classifier to effectively function, the hardware vendor says it’s the software and the software vendor says the LIDAR data are defective. You then find yourself stuck in a blame game with no way to get your work done. Our True View 3D Imaging Sensors/Systems (3DIS®) include full post-processing software from the ingest of raw sensor data (in our case, both laser range data and RGB imagery) to finished or CAD-ready products. In this week’s article I wanted to walk you through the tools that are available in our True View EVO software. I’ll do this in a bit of a workflow oriented manner.
Sensor Post-Processing From raw data to colorized point cloud in LAS format (all the below is accomplished via an easy to use wizard):
Import laser range data, camera images, raw POS data, base station observations, check point locations
Process a refined trajectory using (included with all sensors) Applanix POSPac, directly from within EVO
Geocode the laser data, forming a point cloud
Colorize the point cloud – this uses GeoCue’s unique 3D point tracing algorithm
Quality Check the Data
Use EVO coverage and density measurement tools to ensure you have fully covered the project
Measure geometric accuracy using EVO’s American Society for Photogrammetry and Remote Sensing (ASPRS) compliant accuracy assessment tools
Check strip to strip alignment using a profile view with draped lines color-coded by flight swath
The next steps in the workflow depend on your planned delivery. For example, if you are doing stockpile volumetrics, it is not necessary to do a complete project ground classification.
Creating a Ground Surface, Export Ground Raster, Contours:
Remove low/high noise (if present) using EVO’s automatic noise filters
Classify ground using our automatic ground classification algorithm
Inspect the ground quality using dynamic (“on-the-fly) contour generation tools
Fix blunders using EVO’s extensive interactive classification tools
Use EVO’s Export Wizard to create a gridded elevation model and/or topographic contours
Simple Stockpile Volumetrics:
Use the automatic toe generation tool to create a 3D feature that designates the stockpile toe. Note that this tool will also automatically classify any overhead structures (such as conveyors) so that their points will not be used in the volume calculation. Figure 1 depicts the use of advanced tools in EVO being used to compute volumetrics.
Clean any errors in the toe boundary using EVO’s feature edit tools
Remove any low/high noise within the toes using EVO’s automatic noise removal tools, constrained to run “by polygon”
Run the volumetric computation tool to complete the operation (this tool encodes the toes with volume attributes and optionally generates “Cut/Fill” images)
Figure 1: Stockpile Volumetrics In EVO
Differential Volumetrics (Cut/Fill, borrow pit) These measurements require identical baseline datums:
Ensure equivalent baselines by extracting check points from one surface in undisturbed areas and measuring in the other
Shift one cloud or the other using EVO transform tools
Use automatic toe creation tools or digitize a toe
Perform the differential computations by using one point cloud as the base and the other as the “hull”
Roof print Extraction (building “footprints”):
Classify ground as detailed above
Classify planar surfaces using EVO’s automatic planar surface extraction algorithm
Optionally create polygons of the extracted planes using EVO’s automatic “point tracing” tools
Rail Extraction (ground classification is not needed prior to using this tool set)
Digitize a short seed near a rail centerline
The automatic rail extraction tool will classify top of rail and, optionally, draw a 3D centerline feature. This feature is attributed with the point to point superelevation between the rails
Transmission Line Classification, Extraction (ground classification is not needed prior to using this tool):
Digitize an approximate centerline for the transmission line corridor
Run the automatic wire extraction tool. This will:
Classify the points that represent wires
Create catenary features that are the best fit to the wires
Create correct topological connections between wire connections at isolators
Label phases and guards
Hydro Enforced Elevation models:
Digitize water boundaries using computed constant elevation tools – EVO supports nested “water (or void) on right” breakline creation for “island within island” digitizing
Classify water points using “classify in polygon” tools
Digitize flowing stream talwegs using “downstream constraint” tools
Digitize wider flowing features using “double line drain” tools
Generate rater elevation models and topographic contours using the above features as “breaklines” (3D elevation model constraints)
Cross-sections (Profiles):
Classify the surface from which profiles will be extracted (usually ground, road)
Digitize or import an alignment (a centerline which can be a circuitous polyline)
Specify profile interval and width
Run the extraction tool
Head-up Digitizing There are a number of tools in EVO that allow you to interactively create 3D features (points, polylines, polygons). Some examples of these tools are:
Import linework from a variety of formats (including an optional spatial reference system transform)
Surface Drape (“drape” features over a triangulated irregular network, TIN, generated from the point cloud)
Automatically set vertices based on closest, mean, lowest, highest point
Adjust existing features (perhaps imported) to the point cloud
Shift data up, down or planimetrically
Vertex by vertex editing tools
Smooth linework (using a Douglas-Peucker algorithm)
Export linework in Shape, DWG or DGN format
Split, merge and so forth
Many more capabilities!
Encroachment:
Create or import the 3D feature of interest (for example, wire catenaries)
Run EVO’s “Classify within distance of” tool to classify encroachment points
Web Viewing:
Use EVO’s export EVO Explorer package to create a view “site” in GeoCue’s Amazon Web Services-hosted True View Reckon
Explorer Package This is a data set you can package as a zip file and deliver to your clients. Your clients can use our free EVO Explorer software to visualize and perform measurements using point cloud, feature and image data:
Use EVO’s Explorer Export tool to create a single, composite zip file of the data you wish to deliver
Your user imports this Explorer Package into EVO Explorer using a simple import command.
Seamless integration with Agisoft Metashape
We have integrated what we call “Metashape for EVO (MfE)” directly into EVO to allow creation of photogrammetric products without ever leaving the EVO environment. This is a very rich set of functionality that allows POS data to be injected into photogrammetric models where correlation cannot solve for image positions. The details of this will be provided in a separate article.
The list above is just a sampling of the many tools in True View EVO. Our development is very active and uses a code base we share with LP360. This means that our UAS customers benefit from all of the traditional tools that we add to LP360 in addition to those specifically designed for UAS mapping.
I encourage you to look at the complete tool set that will be provided by your LIDAR/Camera system vendor. I think you will find that GeoCue offers the most comprehensive solution for UAS-based, fused LIDAR/Imagery in the market. After all, we are in the business of selling production solutions, not just sensors.
留言