Beekeeping With a “Flir One for Android”

I will update this post continually (instead of writing multiple posts that could easily take over this blog) as I explore the capabilities of the Flir One infrared camera device. The updates will appear in descending order. The latest update was posted on December 19th, 2016.

DECEMBER 19, 2016: I know people who are getting much better results with their Flir One than I am. I realize my long rambling post here probably isn’t providing the best information because I’ve more or less taken the point-and-shoot approach. I want to turn this thing on, take a thermal image and immediately see something that’s useful — without having to modify the image later on using another application or program, because I don’t have time for anything else. Easily 95% of the images I get following that approach don’t show me anything that helps me. For people like me who may have day jobs, family responsibilities or other activities that don’t allow them much time to fiddle with something like this, I’m not sure I can recommend the Flir One. It’s too expensive and time-consuming. I’m not saying it doesn’t work, but in my experience so far, it’s definitely not something that produces great results just by plugging it in. It takes time and perhaps some knowledge of thermal imaging to get the most out of it.

DECEMBER 07, 2016: I took the following thermal images with my Flir One today. The automatic settings didn’t give me much to work with, so, again, I tried to adjust the settings on the fly and got various results. Using the Flir Tools app or the software on my computer would probably produce better images, but I don’t have time for that. If I can’t get a half-decent image by using the Flir One as I would my cell phone camera, then it could easily become yet another beekeeping chore that takes up more of my time. I’d rather instant results, not something that requires additional processing afterwards. So I’m aiming for simplicity at the moment even if it means I’m not utilizing the Flir One to its full potential. So… It was about -10°C when I took these pics, and no wind. It felt icy cold. Everything was frozen.

The wrap on this hive is loose on the bottom left.  I can't get a useful thermal reading off it. (Dec. 07, 2016, 10:16am, -10°C.)

The wrap on this hive is loose on the bottom left. I can’t get a useful thermal reading off it. (Dec. 07, 2016, 10:16am, -10°C.)



Although some of the cluster is breaking above the top bars (I noticed it when I added sugar bricks last week), the cluster seems to be concentrated on the right side of the bottom deep, and it's been that way all along. This hive is wrapped and has a moisture quilt. (Dec. 07, 2016, 10:17am, -10°C.)

Although some of the cluster is breaking above the top bars (I noticed it when I added sugar bricks last week), the cluster seems to be concentrated on the right side of the bottom deep, and it’s been that way all along. This hive is wrapped and has a moisture quilt. (Dec. 07, 2016, 10:17am, -10°C.)

These bees were between the middle and top deep for a while, but seem to have moved mostly to the top deep now.  The hive has a piece of hard insulation on top. (Dec. 07, 2016, 10:15am, -10°C.)

These bees were between the middle and top deep for a while, but seem to have moved mostly to the top deep now. The hive has a piece of hard insulation on top — which has nothing to do with the colour profile on this image. (Dec. 07, 2016, 10:15am, -10°C.)

This hive was jam packed with bees going into the fall and seem to ignore the bottom deep.  The bees are breaking over the top bars. Is the cluster so large that it's spread over the top two deeps?  I don't know.  The hive has a pice of hard insulation on top and it's not wrapped, though I think I'll wrap it soon.  (Dec. 07, 2016, 10:17am, -10°C.)

This hive was jam packed with bees going into the fall and the bees seem to ignore the bottom deep. The bees are breaking over the top bars. Is the cluster so large that it’s spread over the top two deeps? I don’t know. The hive has a pice of hard insulation on top and it’s not wrapped, though I think I’ll wrap it soon. (Dec. 07, 2016, 10:17am, -10°C.)

This hive is being blasted by the sun. (Dec. 07, 2016, 10:13am, -10°C.)

This hive is being blasted by the sun. (Dec. 07, 2016, 10:13am, -10°C.)

NOVEMBER 24, 2016: Okay, so the Flir Tools Mobile Android app comes in handy. (I assume it’s also available for iPhones.) I took the following image using the Flir One on my phone and then viewed it through the Flir Tools app on my phone, where I easily enhanced the image to show me the hots spots. The original image taken using automatic settings didn’t look anything like this.

2016-11-24-07-45-43

The resolution of the image is even lower than the original image produced by the Flir One, but whatever, it still works. I still can’t get the Flir One to produce consistently useful images on automatic settings, but once I view the images through the Flir Tools app, it’s not a problem.

Tentative conclusion: The Flir Tools mobile app is essential to get the most from the Flir One, or to even get consistently useful images. That’s assuming what’s shown in the thermal image is what’s actually in the hive. It’s difficult to confirm that without lifting the top off the hive and actually looking inside the hive, which I’m reluctant to do in cold weather. However, the last time I did that, exactly a month ago, the results weren’t too bad.

NOVEMBER 23, 2016 – PART 2: Wrapping the hives seems to reduce the effectiveness of the Flir One thermal images. For instance, what you tell from this image of a wrapped hive?

2016-11-23-07-45-43

Nothing. It’s all lit up. It looks like the whole thing is warm. I can’t tell where the cluster is by looking at the image. I kind of botched the wrap on that one. It’s not as tight to the hive as some of the wraps. It’s pretty much the same with this single-deep hive:

2016-11-23-07-46-32

Yes, there’s obviously more heat coming out of the top entrance, but that’s all I can tell. The wrap is very loose on that hive. And then I turn to this wrapped hive:

2016-11-23-07-47-38

Although the hot spots are dimmer, it’s not too difficult to guess where the cluster is hanging out: the bottom and the ride side of the hive, and up on the right side a bit. (The last time I looked inside this hive, about a month ago, that’s exactly what I saw too.) When I look over at the left side of the hive, I see this:

2016-11-23-07-47-49

No heat signatures on the left side of the hive. Which is what I’ve seen every time I’ve taken a thermal image of this tightly-wrapped hive. I’m not sure why the bees are favouring the right side of the hive, though I think it might have to do with the fact that they get most of their sun in the morning on the right side of the hive.

Tentative conclusion: Hive wrap affects the thermal readings, depending on how tight the wrap is to the hive. The hive that I get the best reading from was wrapped tight. The others, I kind of screw them up and wrinkled the roofing felt a bit. Silvery commercial hive wraps would likely render the Flir One completely useless. As stated in the Flir One Manual, shiny surfaces don’t provide an accurate thermal reading at all.

NOVEMBER 23, 2016 – PART 1: I installed the Flir Tools software on my PC at home and found it to be a bit too much to figure out on my own. But I’ll say this: It’s loaded with all kinds of good stuff that I’m sure anyone who knows anything about thermal imaging will love. I’ve linked to some online YouTube tutorials in the comments that look helpful, but I’m trying to approach this as I imagine many beekeepers with day jobs would: plug it in and go because we don’t have time for much else. If I need to read a giant manual and watch hours of video tutorials and study the history of thermal imaging to find some use for the Flir One, then I know the appeal of the Flir One will drop significantly for many. So for now, I’m passing on the desktop software because I know too many beekeepers who would do the same, who just want to take a picture and be able to get useful information from it immediately.

I’ve been able to get some good images from my Flir One, images I can interpret immediately — but sometimes I get nothing from them. For those photos, I’ve had to install the Flir Tools Mobile app on my Android phone — and I found it easier to use than the desktop software. I could take an image, view it in the app and then fiddle with some simple settings and come up with an image like this:

2016-11-23-07-45-29

It’s not exactly mind blowing, I know, but basically I switched to the original photo and then turned up the thermal imaging so the hottest spots bled through the photo. It might not look pretty (and I probably couldn’t do it again if you asked me to do it at this moment), but I can take a pretty good guess where the cluster is from that image — between the middle and top deep. So… that’s useful.

Tentative conclusion: Someone with more time on their hands than me would probably benefit from exploring both the desktop and mobile version of the Flir Tools software.

NOVEMBER 10, 2016: Installing the Flir Tools desktop software onto my PC could have been more straightforward. So everyone who works at Flir, listen up — here’s how you make it much easier for a casual user. And I say casual user because I doubt anyone using their smartphone to take thermal readings of anything is a hardcore customer. I don’t need, and I don’t want, to provide my non-existent company name or select the industrial application that I want to use the software for, just to download the software. I’m a hobbyist beekeepers who wants to use the “Flir One for Android” to make sure his bees haven’t froze to death over the winter. That’s all. I paid for the device. Give me a download link and a serial number that I enter and that’s it — the software downloads, I install it and it works. The whole process should take about a minute or two, tops. I got the software installed properly, I hope, loaded up a few images, but then realized I was no longer in the mood to figure out how to use it. So I gave up and wrote this update instead. Supposedly the software isn’t too complicated (it looks easy to use in YouTube videos). But we’ll see.

Tentative conclusion: Installing the Flir Tools software is not as simple as installing an app from Google Play — click yes to the permissions, install for ten seconds and off you go. It should be that easy — especially for customers who are using the Flir One on their smartphones and aren’t trying to detect a plutonium leak at a nuclear power plant — but it isn’t that easy. I’m not saying the registration and download procedures are overly convoluted, but enough so to be annoying. I also noticed during the registration process that my device was recognized as a “Flir One for iOS,” even though I have a “Flir One for Android” device.

flir-numbers

Hopefully none of that matters. I’ll test it out later when I’m less annoyed.

NOVEMBER 08, 2016: The automatic settings on the Flir One produce results that at times seem helpful and other times not. To get the most of the Flir One, it may be necessary to install the desktop Flir Tools software for it (if you can figure out what software to download). Here are some images I took this morning around 8:00am with frost on top of the hives and the sun just coming up:

Flir One for Android on automatic settings. It looks like the cluster is in the top box, nothing in the bottom. (Nov. 08, 2016, a cold frosty morning, 8:00am, the sun just coming up.)

Flir One for Android on automatic settings. It looks like the cluster is in the top box, nothing in the bottom. (Nov. 08, 2016, a cold frosty morning, 8:00am, the sun just coming up.)

Automatic settings. The cluster appears to be in the middle deep, which corresponds to my observations of seeing the bees using the bottom entrance, not the top.  (Nov. 08, 2016, a cold frosty morning, 8:00am, the sun just coming up.)

Automatic settings. The cluster appears to be in the middle deep, which corresponds to my observations of seeing the bees using the bottom entrance, not the top. (Nov. 08, 2016, a cold frosty morning, 8:00am, the sun just coming up.)

Automatic settings on a hive wrapped in roofing felt. Not much to get from this image. Perhaps the wrap needs to be pressed tighter to the hive?  (Nov. 08, 2016, a cold frosty morning, 8:00am, the sun just coming up.)

Automatic settings on a hive wrapped in roofing felt. Not much to get from this image. Perhaps the wrap needs to be pressed tighter to the hive? (Nov. 08, 2016, a cold frosty morning, 8:00am, the sun just coming up.)

Automatic settings on a hive also wrapped with roofing felt, but the wrap may be tighter. Or is the cool area on the bottom just loose wrap that's not pressed flush against the hive?  (Nov. 08, 2016, a cold frosty morning, 8:00am, the sun just coming up.)

Automatic settings on a hive also wrapped with roofing felt, but the wrap may be tighter. Or is the cool area on the bottom just loose wrap that’s not pressed flush against the hive? (Nov. 08, 2016, a cold frosty morning, 8:00am, the sun just coming up.)

Here’s what the hive looks like in Human Vision:

Hive wrapped with roofing felt. (Photo taken on Nov. 06, 2016.)

Hive wrapped with roofing felt. (Photo taken on Nov. 06, 2016.)

Here you can see my shadow on this hive, capturing one of the rare times I've worn a baseball cap.  (Nov. 08, 2016, a cold frosty morning, 8:00am, the sun just coming up.)

Here you can see my shadow on this hive, capturing one of the rare times I’ve worn a baseball cap. (Nov. 08, 2016, a cold frosty morning, 8:00am, the sun just coming up.)

Another hive wrapped in roofing felt. Not getting much from this image.  (Nov. 08, 2016, a cold frosty morning, 8:00am, the sun just coming up.)

Another hive wrapped in roofing felt. Not getting much from this image. (Nov. 08, 2016, a cold frosty morning, 8:00am, the sun just coming up.)

The left side of this hive is in shadow.  (Nov. 08, 2016, a cold frosty morning, 8:00am, the sun just coming up.)

The left side of this hive is in shadow. (Nov. 08, 2016, a cold frosty morning, 8:00am, the sun just coming up.)

And the right side getting hit by the morning sun.  Is it really 31°C (88°F) on that hot spot?  (Nov. 08, 2016, a cold frosty morning, 8:00am, the sun just coming up.)

And the right side getting hit by the morning sun. Is it really 31°C (88°F) on that hot spot? (Nov. 08, 2016, a cold frosty morning, 8:00am, the sun just coming up.)

More tentative conclusions: It may be difficult to get useful images off hives that are wrapped or loosely wrapped. And automatic settings are okay, I guess, but the Flir Tools software may be necessary to get the most from the thermal images.

OCTOBER 28, 2016: I’m not sold on the Flir One yet. It will take more experimenting with it for me to recommend it. Here’s one of the images taken from my post about pre-winter hive adjustments that looks great, but I don’t know…

I'd call that a hot spot.

I’d call that a hot spot. Is it the exact location of the cluster? Maybe.

It took some fiddling with the Flir One app to get those hot spots to jump out like that. The image was not created with the default automatic settings.

Tentative conclusions: Anything other than the automatic settings, while eventually producing images that seem okay, probably require too much fooling around for people who aren’t tech-savvy. Most of the images I get from the Flir One (which I haven’t bothered to upload) do not reveal anything useful. There has to be a simple way to fine-tune the sensor (I hope there is), but I haven’t found it yet.

OCTOBER 26, 2016: Out in the dark, I put my hand on the side of my shed…

ir-2016-10-25-22-28-54

…and then removed my hand and took a shot of the heat my hand had left behind:

ir-2016-10-25-22-28-56

I don’t know how I managed to get a thermal reading this sensitive but still couldn’t get a single usable image of my hives.

OCTOBER 25, 2016: I also noticed a huge difference in the images by changing the angle of the camera by the slightest degree. The hive in the following shot, for example, looks thermally cool:

ir-01

But two seconds later at a slightly different angle, the image is not the same:

ir-02

The issue may lie in a setting referred to as “Lock Span.” Here’s a long quote from the instruction manual that explains it (the emphasis are mine):

“Ordinarily, the camera uses an Automatic Gain Control (AGC) process to automatically adjust the image based on the range of temperatures that are in the scene. The camera detects the range or span of temperatures, and maps the temperatures to colour in the colour palette dynamically. As a result, the colour of an object at a certain temperature can vary, depending on the other temperatures in the scene. When measuring temperatures with the Spot Meter, sometimes it may be desirable to temporarily lock the range of temperatures to a certain range, so that it is possible to compare multiple images taken at different times and of different scenes.

That’s easier said than done. I tried locking the temperature to a certain range and even took the photos several hours after sunset so the heat from the hives would be stand out more — and the results were garbage:

What can I tell about this hive from this image?  Not much.

What can I tell about this hive from this image? Not much.

I was out there stumbling in the dark for about 20 minutes (until the Flir battery died), flicking every setting on and off, trying to compare the difference between auto and manual settings, and I didn’t get a single image that I could use. Once I turned off the auto settings, I couldn’t seem to turn them back on again. I had to uninstall and reinstall the app to clear all the settings.

Tentative conclusions: The automatic settings produce images that are the easiest to make sense of, but are not the greatest for side by side comparisons because the same object can look different depending on what other objects are in the shot and how the automatic gain happens to adjust the image. Useful images are hard to produce through the non-intuitive manual settings.

I need to find a YouTube video or some other tutorial to figure this one out. I may also need to learn more about the science behind thermal imaging and how to interpret infrared images. I’ll try out the Flir Tools desktop software too. Flir, by the way, is pronounced Fleer, at least according to the Flir Tools video. I would have never guessed that.

DAY 1 – OCTOBER 24, 2016: I recently purchased an infrared (or thermal) imaging device called a Flir One for my Android smartphone (also available for iPhones) to help me monitor my honey bee colonies in the winter. I view the thermal imaging camera as another gadget in a long list of beekeeping fads reserved principally for people with lots of money to blow, an unnecessary expense for most hobbyist beekeepers. If my Flir One hadn’t been given to me as a birthday gift, I’m not sure I’d fork out nearly $400 Canadian for it. But seeing how I have one, it might be fun to see what it can do. Maybe I’ll be convinced it’s a wonderful thing in the end. I should also point out that I have no previous knowledge of infrared technology, so if it seems like I don’t have a clue what I’m doing, you’d be right about that.

I thought about buying an infrared camera after one of my colonies more or less froze to death last winter. I goofed and left a blank frame of bare foundation in the top of the hive that acted as a barrier that split the cluster in two as the bees came up to access their honey stores. Two small clusters don’t produce as much heat as one large cluster. And so it goes. If I’d had a better sense of the cluster’s position inside the hive, I might have been able to prevent the loss. Specifically, a thermal imaging camera may have revealed that the cluster was split in two, allowing me to take action right away and put the cluster back together. Better beekeeping could have prevented it too, but anyhow…

I think I’ll need to fine-tune the Flir One to get the most of out it (through use of the Flir One Android App). The first set of thermal photos I posted on October 22nd, 2016, were taken using mostly the default settings. I turned on the “Spot Meter” that displays the temperature in the cross-hairs and I switched the temperature reading to Celsius instead of the antiquated Fahrenheit scale. Those images turned out okay, but who knows how precise the readings are? I have nothing to compare them to.

So today when I took more photos, I fooled around with the “Emissivity” setting, which apparently refers to how shiny, smooth or rough an object is and how well it radiates heat. According the Flir One instruction manual (PDF), “glossy or reflective materials like metals tend to be poor emitters.” My bee hives are painted with a semi-gloss paint, so my first thought was, “That’s not good.” But after recording images with the lowest (or roughest) Emissivity setting called “Matte,” and then moving up to the highest (or smoothest) setting called “Glossy,” the differences between the images were zero. There’s no point in posting the photos for comparison because they all look the same.

However, the colour of the object seems to have a greater effect on the thermal reading than the texture of its surface.

[PHOTO REMOVED because what I originally described was not exactly accurate.]

Does the white paint give off a cooler infrared reading? It kind of looks like it does. It makes sense that lighter colours would emit less heat in direct sunshine than darker colours. But does that apply to objects under indirect daylight as well? Maybe.

The bottom deep in the next photo is painted white, and shows up cooler in the image, whereas the top deep is dark green and appears warmer in the image (though most of the bees could be in the top of the hive giving off more heat):

ir-white-bottom-box

A tentative conclusion: Different colour hive components give off different thermal readings. A uniform colour for all the hive components might work better.

4 thoughts on “Beekeeping With a “Flir One for Android”

  1. This is the same device I’ve been given to try out. My hives aren’t painted so will be interesting to see how my results compare to yours.

  2. My hives aren’t wrapped yet either. I wonder how that’ll change things… if I wrap them.

    I’ve taken more photos with it. The results are getting better, though I can’t seem to adjust the thermal sensitivity. The way I’m winging is, I turn off the auto-gain, then point the camera at something hot or cold (my body or the ground) and then press the recalibrate button, so it sets a new baseline, or a new minimum, for the thermal signature to show up. But there’s nothing exact about that and there doesn’t seem to be a way to calibrate that exactly… if I’m making sense.

    I haven’t downloaded the Flir Tools software yet.

  3. There doesn’t appear to be any precise way to calibrate the Flir One. It’s either on automatic settings, which half the time don’t produce images that show much of anything, or a manual calibration settings, which I described in my last comment — and are not at all precise.

    It’d be more useful if I could set the dynamic range of the colours. For instance, the coldest colour might be set for 5°C and the hottest 20°C. Basic manual settings like that don’t seem to be available through the app.

    I haven’t had time to poke around much the with Flir Tools software yet, but what little poking around I have done seems promising. But just looking at the software tires me out. I’m just don’t have the time or the energy to dig into it right now.

    The people at Flir may want to consider a consumer and professional version of the app and the software. A simplified consumer version would be preferable for my needs. That, or they need to provide a instructional video that shows consumers how to get the most out of the software. Just some friendly suggestions.

    If I didn’t have a full-time job and had more time to try to learn the software, I might not have any of this issues. At this rate, I’ll probably have my head around all of it by Xmas.

Comments are closed.