Placing our pictures on the globe

Over Jan Plan I worked with Manny Gimond in Colby's Geographic Information System (GIS) lab. Manny taught me how to georeference our aerial photos using ArcGIS software. We use the software to ultimately assign a latitude/longitude coordinate to each pixel within our aerial photo. Once georeferenced our imagery can then be used by popular GIS software like Google Earth.


The tie lines here show points where pixels in our aerial image match points found in an already georeferenced aerial map. ArcGIS software uses the tie lines we demarcate to georeference our image.


Here's a screenshot of ArcGIS showing our georeferenced aerial photo of Colby's new football field. Soon we'll be able to use our UAV's on board GPS unit to record the place (lat/long coordinates) at which each of our images are taken. In the future this GPS meta-data will help us automate the georeferencing process.

Popham Beach

Over the last two weekends I've been venturing to Popham Beach in Phippsburg, Maine for a geology project. I'm investigating the dune erosion caused by the encroaching Morse River. The Morse River is eating up the sand barrier which houses the foundation for a new bath house constructed in the last few years. I hope to document the erosion with aerial photos captured from our UAV over the spring semester. In doing so we may be be able to evaluate erosion, deposition, and sediment flow rates of Morse River.


A video I captured flying our small V-tail hornet remote control over Morse River. There was no autopilot on board so the flight's a bit shaky...weak stomached beware!


Picnic benches down at the margin of erosion.


The new bath house complex standing ominously in the distance.


View of the Morse River with Tucker for scale.


A video Foster took of the Morse river. This is active Geology!

Waypoint Flight over Colby

Today Foster and I along with a few guests flew a 6 way point path over Colby. We used Google Earth to plan our flight and upload 3 dimensional way points to our autopilot. Our planned path is delineated below in white. Our actual path is shown in pink. We took off at the turf field, flew over Johnson Pond, around Miller Library, and over the football field. Next step is getting our camera to fire images at distance intervals while in flight.

Analytics from Friday's flight

Our autopilot records flight data 5 times per second while in the air. Manny Gimond, the GIS instructor at Colby, helped us manipulate that data in Excel and graph some of our flight metrics vs. time. The graphs depict a time series of autonomous mode sandwiched by two manual modes. During manual modes the plane is either taking off or landing and we are in control through a radio. During autonomous mode the autopilot is in control and attempting to achieve mission targets.

Mission targets:
Flight path - holding circle (picture in our last blog post)
Altitude - 208 meters above sea level
Speed - 45 km/h

Autopilot Adjustments

In order to adjust the way our autopilot controls autonomous flight we tinkered with parameters defined in it SET file. The SET file defines particulars about an aircraft which assists the autopilot in flight. Our goal was to control the planes altitude control more effectively. With help from the attopilot's creator, Dean, and other's at support forums, we decided to change the pitch and roll gains. Higher gains equate to higher physical responses to changes in pitch and roll. We need higher pitch gains in particular because our plane has a large glider like wingspan. After a few flights with different pitch gain adjustments we we're able to control altitude much better in our holding circle. Here's our flight path from google earth. The plane held altitude at around 130 meters above ground.

Google Earth Flight Log

Attopilot records over 70 flight data parameters at 5hz into its "black box" aka log file. Using ground control software we can upload the log file into google earth to visualize the flight track. Here's a 3D screen shot of the flight path in google earth. The white path indicates remote control mode while the red indicates autonomous mode. You can see on the flight path where we retook radio control at the crest of the flight pattern.

First UAV launch

Today Foster and I launched our UAV for the first time out of the Colby lacrosse turf field. We set the plane to fly a circular holding pattern around the field to see how the autopilot controlled the planes stability, altitude, and velocity. We launched the plane in remote controlled mode and switched into autonomous flight mode soon after. The plane flew very stably with a consistent airspeed but continually gained altitude beyond the limit we set in the autopilot. We believe this has something to do with the elevator control surface servo which we'll test out again this week. Here's some videos of our first autonomous flight.