Introduction
The entire Geospatial Field methods class had the opportunity to visit the Litchfield mine in Eau Claire, WI for a data collection excursion and to witness a selection of UAS platforms in use. In this field activity, we were introduced to the practice of Ground Control Points (GCP), and we used a variety of GPS units at these points to demonstrate the range of accuracy between consumer grade and survey grade technology. In addition, we were to compare the different UAS platforms we saw operate, noting its features and considering its pros and cons. Lastly, a topographic survey was demonstrated for us using a total station.The primary focus of this activity was data collection. The processing and analyzation of the data will be done at a later date.
Study Area
Our study area for this activity was Litchfield mine, 44°46'23.3"N 91°34'19.5"W - located in Eau Claire Wisconsin (Figure 1).
Figure 1: Litchfield mine
Litchfield mine is a fluvial aggregate mine. A dredge forms large piles of aggregates that were in the massive pond were formed thousands of years ago from a river carrying glacial material with it. Working at the mine, one must know the volume of the aggregate piles. That is where the total station and 3-D model created by the drones we saw come in.
Methods
Our first task in this field activity was to lay out GCPs throughout the mine while practicing drawing out a map of our study area and the locations of the GCPs. The purpose of the hand drawn map was for reference when we were to visit the GCP again to collect coordinates.
As a group, we carried out 16 GCPs (Figure 2). As a general rule, the more variability in the terrain prompts more GCPs. However, one must be careful not to cluster the GCPs in one spot because that can cause distortion in the final image produced.
The first GCP was laid towards the center of the study area and we made our way through and around the mine, laying out 16 GCPs in total (Figure 3).
As a group, we carried out 16 GCPs (Figure 2). As a general rule, the more variability in the terrain prompts more GCPs. However, one must be careful not to cluster the GCPs in one spot because that can cause distortion in the final image produced.
The first GCP was laid towards the center of the study area and we made our way through and around the mine, laying out 16 GCPs in total (Figure 3).
Figure 2: Students following the professor's lead, carrying the GCPs.
Figure 3: GCP 1.
Once we finished, the class broke up into groups, averaging about 5 students and an instructor. Each group had a survey grade GPS and a commercial grade GPS. The group I was in had the Topcon HiPer HR for its survey grade GPS. Every group had a Bad Elf GPS unit. Every student had a smartphone. We were to collect coordinate points with each of these technologies at a GCP.
We would collect coordinates at the GPCs using the Topcon HiPer Hr first. The red stakes were planted firmly into the ground to hold the system in place (Figure 4). It's important that the central skewer is placed as accurately as possible in the center of the GCP to ensure the accuracy of the image and to minimize error propagation (Figure 5). Coordinates were collected from a central panel on the system.
Figure 4: Demonstration of the set-up of Topcon HiPer HR on a GCP.
Figure 5: Centeral skewer resting on the center of the GCP.
Afterwards, a student used the Bad Elf GPS units to read the coordinates of the center of the GCP (Figure 6). In addition, students used their smartphones to get the coordinates of the GCPs using app such as Google Maps. All students recorded the coordinates generated by the Bad Elf GPS and their smartphone for each GCP.
Figure 6: A Bad Elf GPS units recording the coordinate of GCP 16.
The analysis of the recorded data will be done at a later date. The coordinates available to us currently is the phone and Bad Elf coordinates in decimal degrees (Table 1 and 2).
Table 1: Sample of my group's phone data.
Table 2: Sample of my group's Bad Elf data.
After data collection, we reconvened to discuss another GCP survey technology - Arrowhead Markers. We learned that they are self-supporting, waterproof, and solar-powered. They are made of foam, making them light and durable. An operator just needs to press a button on the GCP and it will record its coordinate location and upload it to the Cloud. They can communicate with the other GCPs and with satellites. They say they have an accuracy of 2 cm, but we were told by Menet Aero that they usually experience a 6 cm accuracy. The points are $500 dollars a piece and you can only buy seven at a time. A downside to the markers is that they don't tell you the coordinates on the markers, you trust that coordinates are successfully sent to the Cloud.
The next section of the field activity involved watching and taking notes on the launches of several UAS flight missions. First off, we were briefed on safety procedures before the flights. We were to stay away from the UAS rotors and behind the catapult once it's set up. Stay at least 20-30 ft. away from the operation. Keep out of the landing area designated by the cones. Be aware of the UAS behavior and your location, as wind can affect flight and launching. A fire extinguisher is always on hand. The UAS batteries are very flammable so only carry two at a time.
Once we were made aware safety procedures, we watched Menet Aero set up a multiple rotor UAS - the DJI Phantom 4 Pro (Figure 7). It uses the 1 inch CMOS with 20 M effective pixels. The operator demonstrated that he manually controls the UAS for a pre-mission flight test to check the altitude, and how to grab the UAS with his hand when it comes in to land when there isn't a landing pad (DJI, 2017).
Figure 7: DJI Phantom 4 Pro.
For the flight mission, a map and route is created and downloaded on a control pad held by the operator (Figure 8). They use DroneDeploy for their operations and stressed having a back-up in case something goes wrong. Complications led the operator to use Pix4D instead.
Figure 8: The operator monitoring the flight mission from the control pad.
The downloaded map makes a polygon of the area to be surveyed and displays the location of the UAS in flight (Figure 9). From the control pad, the operator watches the amount of pictures being taken. The Phantom 4 take pictures based on its altitude and flight. He also keeps an eye on the battery because the battery can drain quickly in the cold and high winds. As a general rule, battery life is 30 minutes for an hour of charging. The operator can also control the camera from the control pad. Point the controller away from you to make right and left direction easier as a helpful tip.
Figure 10: An image of the map downloaded into the control pad.
The drone elevates to an altitude of 250 ft., and it flies at 20 mph. The flight last approximately 16 minutes.
Our next demonstration was a Sensefly Ebee from Topcon Solutions (Figure 11). This is a fix-wing UAS with replaceable foam wings. The system must be sent to D.C. if parts need to be replaced in the body. It uses a 20 mega pixel camera that fits into the top of the aircraft and faces down. Multispectral and thermal cameras are also available for the UAS. The aircraft utilizes RTK.
The aircraft is equipped with a high speed camera that senses the ground when landing, though low texture can cause landing trouble. Additional features include radio telemetry and a GPS antenna. It has "return home" safety features from situation like high wind or a poor GPS signal. A temperature sensor in the aircraft monitors for too high temps. It has a transponder compartment that can notify the operator of its location if it crashes.
To begin the mission, the operator connects to the internet to upload the map. The software can run on any PC. The block generated by the map is 25 acres. An estimated 244 photos will be taken, and the estimated flight time is 11 minutes. We are informed that any UAS flight mission over 60 minutes is considered a military operation.
The Sensefly Ebee motor is started and brought to its home point where it will be launched for a pre-flight check about 700 ft away from us. Once that is completed, the operator returns to home point to hand-launch the aircraft. With the aircraft's nose to the wind, the operator shakes the aircraft three times to initiate the motor. He counts to three and launches the Sensefly Ebee. The aircraft turns off the motor and enters glide-mode to take pictures. This helps keep the camera steady.
Figure 11: Operator preparing the Sensefly Ebee for launch.
While the Sensefly Ebee was in flight, it had a few moments where it spiraled out only to level out again. Eventually, five eagles became very interested in the aircraft and began to close in on it. With the eagles and the aircraft's erratic flight, the operator decided to call it back from the computer. On its way back, the aircraft lost control entirely at 246 ft. and crashed into the Chippewa River. It was eventually retrieved. The computer said the IMU failed.
The next UAS launch was the M600 Pro with a Zenmuse X5 camera. The aircraft also had a GeoSnap Pro attached to it. It is separate from the platform had has its own GPS.
The final aircraft we observed was the C-Astral Bramor with a Sony a6000 camera. This UAS is launched via catapult (Figure 12 and 13). Multiple operators make sure they follow through on the aircraft's checklist to set it up correctly and to avoid dangerous mistakes. This aircraft also launches into the wind. The wind's direction is always tested and monitored with ribbon. The operator also shakes the aircraft three times. The C-Astral Bramor is mounted on the catapult and is prepared for launch (Figure 14). Flight time was approximately 30 minutes.
Figure 12: C-Astral Bramor catapult set up.
Figure 13: C-Astral Bramor aircraft mounted on the catapult.
Figure 14: The operator preparing the C-Astral Bramor for launch
The C-Astral Bramor also had an unfortunate landing. Normally, a parachute deploys for its landing. The parachute did not deploy, and the aircraft crash landed.
The third part of the field activity involved observing a topographic survey with a Topcon Robotic Total Station (Figure 15). The total station faces a gravel pile, and it gathers information about the gravel pile. This total station was a three second instrument, and complete scanning would take an hour.
Figure 15: Topcon Robotic Total Station.
Conclusions
This was my first experience with UAS platforms, so all the information was new and informative to me. My favorite aircraft that I observed would have to be the Phantom 4 Pro. It launches and is controlled through a control pad, which seems relatively easy to handle. The downsides to it are the facts that it is white and thus harder to see in the sky, and you must be careful with the rotors. The Sensefly Ebee had many fascinating equipment in it, like the transponder, GPS antenna, high speed camera, and the availability of multispectral and thermal cameras. The foam wings are replaceable too, which is handy. I also think its cool how the motor shuts off when it takes pictures to steady the camera. However, tight spaces are hard for it to maneuver. It's technological failure and crash landing also didn't leave a good impression. Though I suppose that can happen with any UAS. The C-Astral Bramor, while had the best launch, took a significantly longer time to prepare for launch. I also felt there was unnecessary risk operating this platform. I wonder what the payoffs are for operating this UAS that is more risky than the others.
In another activity, we'll be able to analyze the data we've collected, incorporate the GCP, and compare the images taken by the UAS platforms.
Sources
Imagery (2017) DigitalGlobe, USDA Farm Service Agency, Map data (2017) Google.
DJI. (2017). Phantom
4 Pro: Visionary Intelligence. Elevated Imagination. Retrieved October 10,
2017, from DJI: https://www.dji.com/phantom-4-pro.
No comments:
Post a Comment