ASU LiDAR/ALSM Research
Comparing LiDAR shot counts: GeoEarthScope, ECSZ, and B4 surveys
Ramón Arrowsmith, May-June, 2010
Introduction
A question came up about the actual shot count per square meter or density for various LiDAR topographic surveys that have been done recently. I have written about this before (Analysis of LiDAR shot densities for NCALM Eastern California Shear Zone (ECSZ) survey and B4 and A few notes on the differences between data gathered by the 1233 and 5100 ASLM scanners), but I was not able at the time do do some direct comparisons.
Overlapping LiDAR data
OpenTopography now has LiDAR datasets on line with overlap which can be used to explore this question (see table below).
| Data set name | Source | Acquisition date | Scanner information (AGL, SR)* | Metadata link |
|---|
| NSAF | USGS/NASA | February, 2003 | Terrapoint ALTMS 4036 (1000 m, 64kHz) | NSAF metadata |
| Eastern California Shear Zone (ECSZ) | Oskin/Perg (NSF) | November, 2003 | NCALM Possibly Optech 1233 (600 m, 33kHz) | None available |
 B4 | OSU/USGS/NCALM/UNAVCO/Optech** | May 15-25, 2005 | NCALM Optech ATLM 3100 (600 m, 70kHz) | See references noted** |
| Northern California EarthScope (NoCAL) | EarthScope/NSF | March 21 – April 17, 2007 | NCALM Optech GEMINI (700 m, 125kHz^) | NoCAL EarthScope metadata |
| Southern California EarthScope (SoCAL) | EarthScope/NSF | April 2-24, 2008 | NCALM Optech GEMINI (700 m, 100kHz) | SoCAL EarthScope metadata |
*AGL=Above Ground Level, SR = Scan Rate
**B4 was supported by NSF and references include:
B4 project web pages at Ohio State University
Bevis, M. et.al., 2005. The B4 Project: Scanning the San
Andreas and San Jacinto Fault Zones, Eos Trans. AGU,
86(52), Fall Meet. Suppl., H34B-01
Toth, et al., Extreme precision lidar mapping
Toth, et al., 2006, HIGH-RESOLUTION AIRBORNE LIDAR/CCD MAPPING OF SAN ANDREAS FAULT
Toth, et al., 2007, LIDAR MAPPING SUPPORTING EARTHQUAKE RESEARCH OF THE SAN ANDREAS FAULT
^Most of the NCAL EarthScope data were scanned at 100kHz, but the Gualala section compared here was done at 125kHz because of mostly heavy canopy.
Scan Comparisons
Overview
My effort to compare the scans was not exhaustive. Some more quantitative analyses might be warranted and looking at some of the differences between the scans as being due to actual ground displacement--rather than error--would be a good idea. I looked at 3 different overlaps in my comparison (Figure 1): 1) SoCAL/ECSZ overlap along the Calico Fault, 2) SoCAL/B4 northwest of Parkfield, and 3) NoCAL/NSAF near Point Arena.
Figure 1. Overview map of comparison areas (white rectangles) and active faults of western US.
Method
Overview
To compute the shot densities, I used the OpenTopography Point Cloud and Custom DEM tool to select the areas of interest, download the points, and then compute DEMs and point count maps. An important point is that we use a local binning approach which finds all points within a user specified search radius relative to the user specified grid node spacing for the DEM. This method is in production at OpenTopography and also explained as well as available as a Windows application at this link: http://lidar.asu.edu/points2grid.html. This link provides a simple overview of the idea: Exploration of search radius. The point counts are thus all the points in the search radius which was 1 and 4 m in this study (3.14 m2 and 50.26 m2 respectively). To convert to a shot density per square meter, I divided by the area of the search for each grid node. This tends to smear the results a little, but is representative and basically correct (see the results below). It also is appropriate in the sense that at least I think in terms of DEMs in many cases and so I need to know what resolution is supported by what density of points or points in the search radius for a given DEM resolution.
In addition to displaying the hillshades of the DEMs and image maps of the shot density maps, I computed histograms of the shot densities and also produced detailed maps of representative areas showing the actual points over the shot density map, and computed the average shot density in that rectangular area.
Sample MatLab scripts and functions used in the analysis
Once the OpenTopography computations were completed and the points and resulting grids downloaded, I analyzed the results and visualized the patterns using scripts and functions written in MatLab.
- analyze_points_PtArena_NoCal.m--Main script to load grids and points and produce analyses and graphics presented here. Relies on functions below.
- LoadFile.m--Function to load an Arc Ascii Grid (DEM or Point Counts) into the matrix called Elevation and compute useful things like the nulls count, x and y gradient grids. Most of the action is in reading the header.
- hillshademe.m--Function to compute a hillshade based on arguements of illumination azimuth, zenith, and vertical exaggeration. Needs the elevation gradients in x and y as computed from function LoadFile.m. Originally written by Olaf Zielke.
- plothillshade.m--Given previously loaded Easting and Northing matrices (from LoadFile.m function) and the Hillshds hillshade matrix (from hillshademe.m function), plot a hillshade map.
- show_point_count_map.m--Given previously loaded point count grid and easting and northing matrices (from LoadFile.m function), plot the point count map (inside the Elevation matrix).
- compute_and_display_shot_count_histogram.m--Given a grid of lidar shots per square meter, the number of total histogram breaks, and the maximum shot count to show in the histogram range, function will compute and plot a histogram of points per square meter.
- location_for_count.m--Function to use graphical input from the mouse to get the location of interest and to draw a square with x and y widthofinterest and label it.
- find_and_show_points.m--Given a set of x and y locations of points find and extract the points in the range defined in location_for_count.m function. Plot the points out and report number of points and shot density over the area of interest.
SoCAL/ECSZ overlap along the Calico Fault
I compared the overlap between SoCAL and ECSZ through the entire overlap between the two along the Calico fault just southeast of I-40 (Table and Figures 1 and 2).
Figure 2. Calico fault ECSZ and SoCAL data coverage (see Figure 1 for location).
ECSZ Calico Fault data
Figures 3-5 show the ECSZ Calico Fault data, starting with the entire data set and zooming to the fault zone. Single swaths have shot densities of about 0.7 shots/m2, whereas a double overlap shows 2-3 shots/m2 and a triple 4-5 shots/m2 (see figure 5).
Figure 3. Hillshade of entire ECSZ coverage of Calico Fault and zoom to fault zone.
Figure 4. Shot count per square meter of entire ECSZ coverage of Calico Fault and zoom to fault zone. Histograms at right show the shot count distributions. The mode for the entire coverage is 0.48 shots/m2 whereas the mode for the higher density zone along the fault zone is 2.23 shots/m2.
Figure 5. Shot count per square meter of ECSZ coverage of zoom to Calico Fault zone (upper left is the same area as is shown in figures 3 and 4). The three small boxes are 15x15 m on a side and show the shot count map overlain with the actual lidar returns and then the average shot per square meter is thus calculated independently and corresponds with the radially averaged shot count map.
EarthScope Calico Fault data
Figures 6-8 show the EarthScope Calico Fault data, starting with the overlap area and zooming to the fault zone. Single swaths have shot densities of about 2-3 shots/m2, whereas a double overlap shows 4-5 shots/m2 and a triple more than 6 shots/m2 (see figure 7).
Figure 6. Hillshade of EarthScope coverage of Calico Fault and zoom to fault zone.
Figure 7. Shot count per square meter of EarthScope coverage of Calico Fault and zoom to fault zone. Histograms at right show the shot count distributions. The mode for the entire coverage is 1.77 shots/m2 whereas the mode for the higher density zone along the fault zone is 3.5 shots/m2.
Figure 8. Shot count per square meter of EarthScope coverage of zoom to Calico Fault zone (upper left is the same area as is shown in figures 6 and 7). The three small boxes are 15x15 m on a side and show the shot count map overlain with the actual lidar returns and then the average shot per square meter is thus calculated independently and corresponds with the radially averaged shot count map.
NoCAL/B4 overlap along the San Andreas Fault northwest of Parkfield
I compared the overlap between NoCAL and B4 through the overlap between the two along the San Andreas Fault just northwest of Parkfield (Table and Figures 1 and 9).
B4 data |
NoCAL data |
B4 and NoCAL data |
Figure 9. B4 and NoCAL data coverage (see Figure 1 for location).
B4 data
Figures 10-12 show the B4 data starting with the northwestern 6 km of the data set and zooming to the comparison zone. Single swaths have shot densities of about 2 shots/m2 whereas a double overlap shows 4 shots/m2 and a triple 6 shots/m2 (see figure 7).
Figure 10. Hillshade of B4 coverage of San Andreas Fault northwest of Parkfield and zoom to comparison zone (see figure 1 for location).
Figure 11. Shot count per square meter of B4 coverage of San Andreas Fault northwest of Parkfield and zoom to comparison zone. Histograms at right show the shot count distributions. Note the clear bimodal pattern with the lower mode at 2 shots/m2 contributed by the fringing single swath coverage and the second being double coverage at about 4 shots/m2.
Figure 12. Shot count per square meter of B4 coverage in comparison zone (upper left is the same area as is shown in figures 10 and 11). The three small boxes are 15x15 m on a side and show the shot count map overlain with the actual lidar returns and then the average shot per square meter is thus calculated independently and corresponds with the radially averaged shot count map.
NCAL EarthScope data
Figures 13-15 show the NoCAL data starting with the southeastern several km of the data set and zooming to the comparison zone. Single swaths have shot densities of about 2.2 shots/m2 whereas a double overlap shows 5 shots/m2 and a triple 7+ shots/m2 (see figure 7).
Figure 13. Hillshade of NoCAL coverage of San Andreas Fault northwest of Parkfield and zoom to comparison zone (see figure 1 for location).
Figure 14. Shot count per square meter of NoCAL coverage of San Andreas Fault northwest of Parkfield and zoom to comparison zone. Histograms at right show the shot count distributions. Both histograms are dominated by the double swath coverage through most of the area with 4-5 shots/m2. Shot densities as high as 15 shots/m2 are in areas of returns though vegetation in the southwest and maybe highre elevation (lower AGL(?)).
Figure 15. Shot count per square meter of NoCAL coverage in comparison zone (upper left is the same area as is shown in figures 13 and 14). The three small boxes are 15x15 m on a side and show the shot count map overlain with the actual lidar returns and then the average shot per square meter is thus calculated independently and corresponds with the radially averaged shot count map.
NSAF/NoCAL overlap along the San Andreas Fault east of Point Arena
Finally, I compared the overlap between NSAF and NoCAL through the overlap between the two along the San Andreas Fault east of Point Arena (Table and Figures 1 and 16).
NSAF data |
NoCAL data |
NSAF and NoCAL data |
Figure 16. NSAF and NoCAL data coverage (see Figure 1 for location).
NSAF data
Figures 17-19 show the NSAF data along the San Andreas Fault directly east of the town of Point Arena. These older and lower pulse rate data (see Table) are in feet (California State Plane). I think I still got the conversions to metric correct. The flight pattern is evident in the shot density map (Figure 18) and I have chosen an area that spans the split in the single directed overlapping swaths and the second, higher shot density, zone with the additional N-S flight lines. The stated single swath goal for the survey was 1 shot/m2 (NSAF metadata). Figures 18 and 19 show that this was achieved. The double overlap zones have 1.8 shots/m2 and the multi-overlap (3 or 4x) shows 2.6 shots/m2 (Figure 19). Few laser shots are returned from the Garcia River.
Figure 17. Hillshade of NSAF coverage of San Andreas Fault east of Point Arena and zoom to comparison zone (see figure 1 for location).
Figure 18. Shot count per square meter of NSAF coverage of San Andreas Fault east of Point Arena and zoom to comparison zone. Histograms at right show the shot count distributions. Note the clear bimodal pattern coming from the single and double flight coverages (both have overlapping swaths).
Figure 19. Shot count per square meter of NSAF coverage in comparison zone (upper left is the same area as is shown in figures 17 and 18). The three small boxes are 15x15 m on a side and show the shot count map overlain with the actual lidar returns and then the average shot per square meter is thus calculated independently and corresponds with the radially averaged shot count map.
Acknowledgements
The idea for this study came from a question by Mike Oskin. Thanks to Chris Crosby for edits.
Dataset acknowledgements:
- The NSAF airborne laser swath mapping data were acquired in support of collaborative research by members of the U.S. Geological Survey (USGS) and the National Aeronautics and Space Administration (NASA), with funding provided by NASA's Earth Surface and Interior Focus Area. The data were acquired and processed by TerraPoint, LLC under contract to NASA's Stennis Space Center. The data are in the public domain with no restrictions on their use.
- Eastern California Shear Zone: LiDAR data acquired by the National Center for Airborne Laser Mapping (NCALM) on behalf of Dr. Mike Oskin (UNC) and Dr. Lesley Perg (U of M) as part of their NSF project on fault systems in the Eastern California Shear Zone.
- The B4 project (http://www.earthsciences.osu.edu/b4) created an unprecedentedly accurate surface model along the San Andreas and San Jacinto Faults in southern California that enabled the research reported here. It was supported by the U. S. National Science Foundation and led by Ohio State University and the U. S. Geological Survey. The National Center for Airborne Laser Mapping performed the airborne data acquisition and laser data processing. Optech International generously contributed use of the ALTM3100 laser scanner system. UNAVCO and SCIGN assisted in GPS ground control and continuous high rate GPS data acquisition. A group of volunteers from USGS, UCSD, UCLA, Caltech and private industry, as well as gracious landowners along the fault zones, also made the project possible.
- SoCAL and NoCAL Earthscope data: This material is based on services provided by the Plate Boundary Observatory operated by UNAVCO for EarthScope (http://www.earthscope.org) and supported by the National Science Foundation (No. EAR-0350028 and EAR-0732947). LiDAR data were acquired by the National Center for Airborne Laser Mapping.
Data Access Acknowledgement: This material is based on processing services provided by the OpenTopography Facility with support from the National Science Foundation under NSF Award Numbers 0930731 & 0930643.
Last modified: June 19, 2010
