A lot has happened in the last few weeks, so I haven’t really had time to build a kite or really explore that possibility any further. In the mean time, I’ve worked on UAV possibilities we can look into, specifically UAV designs which promote longer flight times. Most of these have been fixed wing designs, which tend to be less of a drain on batteries. One particular inspiration is the Silent Falcon drone, which is designed for LiDAR work, and has 5-12 hour life depending on wing configurations. Building such an aircraft would mean getting carbon fiber moulding equipment in the shop, which I think will be useful no matter what drone we decide to build. Carbon fiber is pretty much the standard for quality UAV solutions, and it will make sense to make as much out of composite materials as we can.

http://www.silentfalconuas.com/Silent-Falcon.html

This week, I spoke to a “balloonist” whose hobby of hot air ballooning will hopefully inform us more on the idea of using hot air balloons, and the kinds of things we should be thinking about. Though it’s not the same thing (he balloons at a scale of hundreds of pounds) there’s tons to learn from his informative e-mail. This advice will be critical when we consider the design of our balloon. His email is pasted below.

”

This week I stopped working on the alcohol burner (at Charlie’s insistence) and started to look for Sterno. I did not find Sterno at Marsh or Tru Value, so someone else might have to pick that up. Once we have that, I’m hoping to use a lightweight plastic or balsa for a the frame for the envelope, though it remains to be seen if either is can be light and strong enough (I’m counting on balsa) .  Meanwhile I started researching potential DIY kite designs from which we could do our thermal mapping. I like the kite idea because it can be used when the hot air balloon cannot (because it’s too windy) giving us more flexibility over when and where we do our thermal mapping. The designs I’m looking at right now are box kites, due to their relatively simple design, ease of construction, strength and stability in the air. I hope to start building a kite out of dowels and TyVek this week, and if it’s stable, I’ll scale those designs to accommodate the thermal camera.

During these weeks I explored the possibilities of using hot air balloons. The trick was, how do I generate enough heat to get lift the thermal camera? The answer is, a LOT. We are going to need to not only generate a lot of heat, we are going to need to have a huge envelope which holds this hot air. I’ve built a couple alcohol burners already, which worked OK. Early models exploded from vapour build-up, and later models had too little pressure to stay lit. Calibrating a burner to burn for long enough while still holding together will be quite the trick. In the mean time, I’ve been thinking: hot air balloons usually float on the calmest days. If Skalines is windy, is this the best approach?

During this week, I actually got to use the thermal camera! It’s super cool to just look at the thermal output from things I would never have imagined was detectable. This includes tracking footprints from the thermal signature left by our feet, the heat given off by the switch panels in CST, tracking living things such as squirrels and people, and trying to identify factors that effect the thermal signature of humans. Thick clothing seems to obfuscate heat, making other places seem to give off more heat. For example when my friend was without a coat, their face registered as 87 degrees F (on the surface) and torso around 82 F, whereas with a coat, torso was practically indistinguishable from ambient temperature (Ok maybe +10, but my point being that coats reduced heat output. I forgot to record this data precisely.), and their head warmed up to 92 +-2 degrees Fahrenheit. I’m still unclear on which birds exactly we will study, and will probably leave that decision to the ecologists in Iceland and perhaps Erin (if she has a preference). Once we know that, it might be possible to more accurately model the birds with maybe a local animal with similar insulative properties and heat output, but that remains to be seen…

In the mean time, I’ve tried to track squirrels with only a thermal camera and my success was very limited. The first problem was that it is hard to find a fixed focus at which the squirrel appears sharp. This might be fixed with the balloon-tether idea which would yield a fixed height from the ground and thus static focal range. However most pressing was the battery life on the camera. I had trouble getting 5 minutes out of the device. If this is standard, we might need to design a new battery, and if it’s a result of slow degradation of the battery, maybe we should purchase a new one. Next week I’ll be sure to post pics of us looking for small animals and how well that works!

Apparently Erin and Charlie are talking to the Maintenance Dept. here at Earlham College about using their thermal camera. This would be an ideal solution, because it removes a huge amount of reinventing of the wheel. It would possibly take decades to get to the level of refinement found in a real thermal camera. In the mean time, I’m mostly waiting on info from the ecologists in Iceland who will tell us which birds we can actually study in Skalines. Until we have an idea of what birds to study, I’m not sure what kind of progress I can really make.

In this week I explored more options for thermal imaging. None of these seemed particularly promising, many of these are either expensive, inaccurate, slow, or the range is limited. So far the thermal sensor options seem to be a “pick any two” of the following options: range, accuracy, cost. Unfortunately, because picking up the thermal signature of a small animal at a sizeable range requires a significant amount of all three, there’s either going to need to be a massive engineering breakthrough on my part, or a massive budget. The same problems persist for other parts of the camera, for example, the optics. Focusing long wave IR requires exotic materials which cost an enormous amount of money to procure individually or manufacture.

My conclusion is that thermal imaging is a field with an extremely high barrier to entry. I doubt that I have the skills or expertise to surmount the difficulties I’ve touched on, and I hope that another solution will present itself.

I was kindly reminded today by Tara to post some of my research so far.

My plan so far, to use a non-contact thermometer to better identify nesting sites, has been primarily based on a paper summarizing the use of thermal cameras to identify nesting sites. Using a Thermographic Imager to Find Nests of Grassland Birds by Edward W. Galligan, George S. Bakken and Steven L. Lima appeared in Wildlife Society Bulletin in Fall, 2003.

This paper, in summary, finds that thermal cameras cannot discern nests from any considerable distance, and found that they were useful in tandem with rope-dragging as a way of confirming nesting sites after potential nests had been previously identified. This is useful for my approach in Iceland, where years of rope-dragging will hopefully mean potential sites have been identified can we can then focus on that data rather than mapping massive swaths of terrain with painstaking detail. My primary ambition is to improve on the technique described in the paper by causing reducing nest disturbances caused by rope dragging.

Beginning this week, I plan to first test the sensitivity and capabilities of a non-contact thermometer. If this is successful, I will start looking for a potential thermopile to use in an array in order to detect heat more quickly. This array will probably covering a 360 degrees view, though much of the configuration depends on the results of our testing with the non-contact thermometer and studies into cost and effectiveness.

Possible variables include:
At what range are these thermometers accurate and fast? Should we be scanning for heat from a UAV or the ground? If we decide to scan from the air (UAV), should we use a quadrocopter or build a RC plane? What angle should we be scanning at? Should there be redundancy or overlap in the thermopiles? Should the sensor array be moving? How fast could it be scanning?

There are many other considerations, but I hope that my testing this week will reveal answers to some of my questions. Hopefully by the end of this week I will have a general idea of how the sensor array will be designed.