- StudyBlue
- North Carolina
- North Carolina State University
- Geographic Information Systems
- Geographic Information Systems 582
- Mitasova
- bmscully.assign5.PDF

Brandan S.

Advertisement

Brandan Scully MEA 592 ? GIS Analysis & Modeling Dr. Mitasova Task The assigned task was to explore methods for importing LIDAR data different methods in GRASS and ArcGIS Approach: Figure 1 shows bare earth LIDAR points same data is shown in Figure 2, sampled at 6 meters. Figure 3 depicts multi-return LIDAR data sampled at 6 meters using the ?range? binning method was used to sample the multi overlaid with vector data representing lakes, streams and roads. An orthophoto of the area of interest is displayed in Figures 5 and 6. Bare earth (Figure 5), and multi return (Figure 6) data points in the range 0.5 The bare earth and multi-return data used to produce the first four figures is displayed together in Figure 7. Here, visualization is in NVIZ. Bare earth data is brown. First return data is green. Data is sampled at 1 meter resolution. Figures 1 through 7 were created in GRASS. Similar functions were explored in ArcGIS. Figure 8 is an ArcGIS image depicting deviation?. Using the attribute table counter it was found that there were 319214 points in the bare earth file. 9/29/09 Assignment 5 at multiple resolutions and using . sampled at 2 meter resolution using the ?n? binning method. The ?mean? binning method. The -return data depicted in Figure 4. Here, the data is -20 meters are shown over the orthophoto. bare earth LIDAR data sampled at 6 meters using ?standard - Figure 9 depicts interpolation of low resolution DEM and DSM using ?radial basis function? method and ?spline with tension? kernel method in ArcGIS. Elevation contours were made from the raster ?elev? at 10 and 3 meters, as depicted in Figure 10. Also depicted are multi return data with the return value of 2. Discussion Figures 1 and 2 show 2 and 6 meter resolution of the same LIDAR data set, respectively. The ?n? method counts the number of points within the grid. As one might expect, the larger 6 meter grids contain a greater number of points than the 2 meter grids. Figure 3 shows the average elevation of multi-return points in each 6x6 grid cell. This is going to be reasonably accurate in any area where all returns are of similar elevation. In areas where return times differ significantly, average elevation should be lower than the actual first surface. White spaces indicate areas where no LIDAR return value is present. This commonly occurs over calm water where the reflected signal is not return to the source. Figure 4 is showing us multi-return data where the range in elevations is between 0.5 and 20 meters. Essentially, this should be highlighting elevations that are not bare earth elevations (the blue areas). The range is created by differing return times in LIDAR data, caused by interfering surfaces, such as vegetation. This is highlighted in Figures 5 and 6. There maximum range in the bare earth data is below 2 meters, whereas there is much more range in the multi-return set, especially in areas of significant vegetation. At 6 meter resolution, it appears that the surface variation caused by things like road beds and some drainage features are lost from what was visible at 1 meter visualized in 3 dimensions, shown in Figure 7. Figure 7 is displaying the bare earth surface and the vegetative surface of the data points from our sample in 3 dimensions. Using the cutting plane, it was possible to see the difference in elevation in the tree stands, much like in the first assignment. Figure 8 is an ArcGIS image depicting the data sampled at 6 meters using the ?standard deviation? binning method. This should behave similarly to the range method in that high standard deviation in data point returns is caused by interference at the scanned surface. The lighter, high standard deviation areas roughly correspond with the blues of Figure 4. By comparing the yellow-brown color ramp of Figure 9 with the elevation contours extracted from raster elev in Figure 10, it can be seen that the spline with tension kernel function produces a close approximation of the elevation surface contours. Figure 10 also depicts the ?return=2? values from the multi-return data set. We should expect these returns where there is some interference between the first surface and bare earth. The GRASS exercise for displaying multi-return points filtered by return value produced an error ?database connection not defined?. The entire multi-return data set was available but could not be filtered. I was unable to create a low resolution DEM to check imported data from ?elevlidrural_mrpts? because there was no ?Z? attribute column. Conclusion: Given the density and accuracy of LIDAR data, it appears that using it to generate surfaces can provide very detailed images. Crisp, detailed images appear to result where resolution is close to point spacing. Sampling at lower resolution appears to smooth details out of the surface. This may be desirable in some instances, where fine detail is not important, and might increase processing speed in large data sets. Both GRASS and ArcGIS provide a robust set of tools for carrying out a variety of analyses. ScullyBM Microsoft Word - bmscully.assign5.docx

Advertisement

"The semester I found StudyBlue, I went from a 2.8 to a 3.8, and graduated with honors!"

Jennifer Colorado School of Mines
StudyBlue is not sponsored or endorsed by any college, university, or instructor.

© 2015 StudyBlue Inc. All rights reserved.

© 2015 StudyBlue Inc. All rights reserved.