Introduction
This field activity will discuss the use of
Pix4Dmapper software in mapping UAS data with Ground Control Points
(GCPs).Pix4Dmapper software allows the user to process drone data with or
without GCPs. Pix4Dmapper strongly recommends the usage of GCPs when combining
multiple different areal images such as nadir (0º) or oblique(45º-90º).
Because of the angle difference in images having GCPs will allow for proper
adjustment of the different set of images that will relay accurate data. Field
Activity #11 discussed the processing of UAS data without GCPs. This field
activity will focus on processing data with GCPs and compare the data accuracy
between the two. The GCPs data for this field activity was gathered with a
Topcon HiPer on 9/30/2016 provided from the University of Wisconsin-Eau Claire.
The area of study is shown in Figure 1.
 |
| Figure 1. Litchfield; area of study |
Methods
The steps of the UAS initial set up for processing of data with
GCPs was the same without GCPs. Referencing Field Activity #11, it goes through
the steps. For the sake of saving time the file that was created from the
initial project setup was saved under a different name. Using this file and
going under the Project Tab-->
GCP MTP manager--> import GCPs. Once the GCPs were imported they
had to be attached to images. To do this Basic Editor was used and going through the 222 images a
couple of images per GCPs were manually tagged. This was able to be done
because each GCPs platform was spray painted with a number. Figure 2. Shows the
GCPs used. Using the number in correspondant to the Label of GCP allowed for
this process to be done. Also, since field notes were taken on the day of this
field excursion a mental map was created when placing the GCP for reference
when flying the drones.
 |
| Figure 2. GCP |
16 GCPs were taken at the Litchfield mine. So every
GCP had a couple of images tagged to it. Not all GCPs were tagged because the
next step was to Process--> Reoptimize. This basically allowed the data to update the points
that were manually tagged. Then going back to Project Tab-->
GCP MTP manager and clicking rayCloud Editor all of the
images were brought in for each GCP but was not tagged. Figure 3. Shows the
GCPs in correspondent to the labels that were not tagged.
 |
| Figure 3. GCPs not tagged in raycloud editor |
After tagging all of the correct GCPs to the labels the data was
then processed through all of the steps. This then created a geotiff that could
be opened as a raster in ArcMap.
Results/Discussion
Figure 4. Figure 4. Shows the Orthomosaic image of Litchfield with
16 GCPs. The Orthomosaic with & w/o GCPs differed from each other. This
could have been just from the processing of stitching the areal images together
individually. Since an overlay of the images would just come up as a blurred
image it was withheld from the results.
 |
| Figure 4. Orthomosaic with GCPs of Litchfield Mine |
Figure 5. Showing the DSMs with & W/O GCPs. The difference can
be seen with the elevation of each DSM. The one on the right is w/o GCPs and
the one on the left has GCPs. The difference in elevation is quite astounding
given just 16 GCPs to go off of in an area of roughly forty acres.
 |
| Figure 5. DSMs with & w/o GCPs of Litchfield Mine |
Conclusion
Referring Figure 5. Data with GCPs is much more accurate when
using drone imagery to survey property and products. The GCPs DSM of Litchfield
mine accurately shows the elevation that can be used to measure the mounds of
sand to estimate the value the mine has. This is very helpful in any industry
and can be applied to others for further use.
Reference
https://pix4d.com/product/pix4dmapper-photogrammetry-software/#
https://pix4d.com/product/pix4dmapper-photogrammetry-software/#
