Georeferencing

          In order to better understand how we can get an image projected a certain way, we can

use GIS. Georeferencing is used to apply coordinates to a map, an aerial photographs, or digital image. With this technique, we can use real life maps in ArcMap and overlay different layers, like streets or buildings to answer spatial analysis questions or get directions. In order for learning purposes, we did an assignment to georeference an image of the UCLA campus.

First, we needed to find coordinates of certain points on the map. We divided the map up into five sections and divided the class into two groups. Each group assigned a few people to go to each section to obtain coordinates of specific points in their section. My group, in team 2, went to section 2, which is the general area of “South Campus.” We collected UTM coordinates from five different locations: the northwest corner of parking lot 2, the southwest corner of lot 2, the southwest corner of another lot, the bruin bear, and the center of the circle near Jan’s steps. When standing in those locations, we used the GPS device to locate our UTM coordinates and recorded them. Then we emailed the 5 coordinates and put together a compilation of all the coordinates collected.

         I then went into ArcMap and georeferenced the points to apply them to the image. I used the georeference tool found in the toolbar. I zoomed into each of the 24 GCP’s that my team located and found the column and row that the location was at. When I found it, I clicked the area and the right clicked it to add the control points. I did this for each of the 24 points. I then checked my points and deleted the points with the highest residuals. As a result, I had a map that had 15 points that all had a RMS error around 8 as shown below.I then rectified the georeferenced data as a tif file. However, after much frustration, it took me awhile, and the help of some friends, to figure out how to project the image and the streets correctly. With their help, I was able to figure out how to get the map to project what I wanted to see. The resulting map has the georeferenced image of UCLA, a street layer, and the GCP’s with graduating symbols showing the residual size.

          From this assignment, I have come to find the discrepancies between GIS and GPS. First of all, when using the GPS locator, it was not precisely accurate. It did not show the locations with decimal points and it therefore rounds the area to the nearest northing and easting. This allows for a large margin of error. This means that a precise location of the Bruin Bear or the corner of parking structure 2 may not be entirely accurate to begin with and therefore could cause misalignment when projecting. Some of the points could be meters off. Furthermore, There were many errors when adding the data into the georeferencing tool. I did not know that I did not need to the “11” in the x data row when entering the coordinates, so therefore I was having problems with projecting the streets and data points. Also, what was really confusing was that there was no description of where the locations were coordinated. It would have been a lot clearer if there was an additional column in the excel file that said something like “northeast corner of drake stadium” or the like, because knowing where the actual point is supposed to go is always helpful.

          Furthermore, there are some irregularities with collecting the data. Some groups did a really good job a collecting data, and I ended up using most of their points. But the group that went to section 5, must have had a hard time because I only used one of their GCP’s. This may be because the landscape of the campus up in that section is very hilly and elevated and therefore the group’s coordinates are less accurate. However, if I were to get a more accurate point, then my maps RMS error would be lower. Also, because the points were generally located near the middle of campus, the accuracy is more precise in that area. If you look at the streets layer, the red lines (the streets) line up more precisely on campus. However, if you look outside of campus in the residential area, the streets are not aligned as well. This is because there are no coordinate points outside of campus and therefore those areas of the street layer do not line up exactly with the image. However, my image was pretty accurate when comparing them to other images my friends showed me.

          In conclusion, GIS and GPS can work together to georeference places on an image and apply real coordinates to them. This can be very helpful when looking at images and applying them to real life situations like finding directions. However, as I have come to find, there is some uncertainty and error when using GIS and GPS because some technologies may not be as accurate or the picture may be pixilated. However, when done correctly, we can have georeferenced maps from any picture we obtain if we can just get enough accurate information.

Sources: http://gis.ats.ucla.edu/mapshare/

By

Chelsea Kemp

May 3, 2011