Introduction:
For this week’s project we conducted a field survey
using a conventional method. To gather
our survey data we used a hand held laser that records both the distance from
an object and the azimuth in degrees.
Although there are more accurate methods for gathering this data, you
cannot always rely on expensive technology as it often fails. The device that we used is called a TruPulse
360B manufactured by Laser Technology Inc. This model provides a number of
measurement features such as slope distance, inclination, azimuthal direction,
and can be synced to data collection software remotely. During the first part
of this lab we began familiarizing ourselves with the equipment and processes
involved in importing the data into ArcGIS. Later, my colleague and I moved on
to survey a 50 meter area within Randall Park.
Methods:
Our class began with a short lesson on the equipment
we would be using and the type of data we would be collecting. Two surveying methods
were used in our research, a standard compass and a sonar range finder, and a
laser range finder with a built in compass. Our objectives were to gather point
data using bearing and distance. Neither
of these methods automatically adjusts for the declination angle at your
location. Before surveying any plot of
land you must be sure to compensate for the difference between true north and
magnetic north. Luckily for us, Eau
Claire is nearly in line with the true north and magnetic north convergence
line. We have approximately one half of
a degree of difference making it quite irrelevant when looking at a small 50
meter plot.
After my partner and I were confident with handling
the equipment, we gathered some data points and imported them into the
GIS. This proved to be a rather
difficult process and the software was quite temperamental. It was important to determine the starting
location from where you were gathering your data. We used a base map within ArcGIS to determine
our location being sure to give it an accurate projection. Once our starting point was determined we
were able enter it into our data table so that our azimuth and distance
recordings were referenced to that point.
Figure 1 shows a sample of test points that were taken in the parking
lot behind Phillips hall. It’s important to note that one of our points was not
accurately represented within the GIS.
It is always important to check your data’s validity and this would
likely have gone unnoticed had we taken a larger sample of points. The long line extending to the west into the
parking lot was supposed to end at the small building about 60 feet to the
north. Since the distance is accurately represented
our azimuth recording must have been wrong.
Figure 1: Sample survey points. Notice the left vertices fall approximately 60 feet south of the actual feature being recorded. |
After our preliminary survey was done Nick and I
moved to Randall Park to conduct our independent survey. We began by locating an easily identifiable
node to record our data. We determined that the sidewalk corner would be easy
to distinguish on a projected aerial photo. Our next step was to measure out
our 50 meter plot (figure 2). We decided
to record simple nominal data on what type of feature was being recorded (i.e.
tree, fire hydrant, stop sign). The
features falling within our measured plot were recorded on a table to be
transferred to excel (figure 3).
Figure 2: Measuring our 50 meter plot. We ended up recording features outside of 50 meters to have a larger sample size. |
Once all of our data was collected we went on to
importing it into the GIS. This step
went much more quickly compared to that of our preliminary survey. After adding
the table and exporting it into a geodatabase, we ran the bearing distance to
line tool. In figure 4 you can see our data represented as lines extending from
a node. This shows the azimuthal
direction of our data as well as the distance represented by the length of the
lines.
Figure 4: Azimuth angles and Distances imported into ArcGIS with an aerial base map. |
The next step is to convert the line vertices to
points. This tool can be found within ArcToolbox under data management tools,
features, and feature vertices to points. Once this tool is run you will now
have feature points for the data you collected. After overlaying an aerial
photograph you can compare your surveyed features to what is seen in the
image. In figure 5 you can notice that
our points fall relatively near the actual features. A higher quality image would make the image interpretation
more clear for example, it is difficult to distinguish a tree from a light post
with that low of resolution.
Figure 5: Our data points overlayed on an aerial image. Feature points can be clicked on to view their identification. |
Discussion:
This exercise provided a relatively simple method of
surveying that can be conducted anywhere.
The technology involved was easy to use and the results we gathered were
surprisingly accurate. There were a few
points that didn’t quite fall where they were supposed to. This was likely do
to some form of human error such as recording the wrong distance. The actual
field we were trying to record was too simple. I would have liked to
have more fields such as trunk diameter or the tree species. Including this data is what would separate our
field study from simply editing point features within the GIS.