FLIR thermal camera with Propeller? It works!
Rayman
Posts: 14,651
Update: Got it working 
(See my posts below)
Anybody already do this? I did a search, but couldn't find anything.
Finally had a good excuse to buy a couple of these FLIR thermal camera modules.
There appears to be plenty of documentation on how to use these.
Looks pretty straightforward with just one way SPI and optional I2C for changing settings.
Think I'll start with just using the Prop to send the data to PC over USB to display and record.
Then, might try displaying on LCD display and saving to uSD, so can be independent system.
The NTSC output of Prop might be interesting for this. I think the Prop has enough memory to buffer the image.
The color palette might be enough to do a decent job.
They are also at Digikey now, but I also got one from Sparkfun since they provide a lot of documentation.
(See my posts below)Anybody already do this? I did a search, but couldn't find anything.
Finally had a good excuse to buy a couple of these FLIR thermal camera modules.
There appears to be plenty of documentation on how to use these.
Looks pretty straightforward with just one way SPI and optional I2C for changing settings.
Think I'll start with just using the Prop to send the data to PC over USB to display and record.
Then, might try displaying on LCD display and saving to uSD, so can be independent system.
The NTSC output of Prop might be interesting for this. I think the Prop has enough memory to buffer the image.
The color palette might be enough to do a decent job.
They are also at Digikey now, but I also got one from Sparkfun since they provide a lot of documentation.
Comments
It's very expensive, or I would have gotten one when they first came out...
The price of the Flir One add-on for the iphone 5 is down to less than $150 on Amazon now. I wonder if it could be hacked in order to avoid the $250 price of the sensor breakout alone? No documentation to be sure. The downward price trend may continue, a consequence of tying a product so tightly to an iphone that is sure to mutate in the blink of an eye.
Admittedly it is entry level, but one thought I have on the subject is that this camera blends a visible image with the infrared image so you can tell what you are looking at. It also ranges the temperature, setting the low and the high based upon what it is seeing at the moment. It will report the temperature in the crosshairs also.
It is a nice little piece of test equipment that I find new uses for every single day, but I am wondering how hard it will be to program any useful features, and how processor-intensive it will be to run it.
From the developer apps page, a seasonal must have!? "Zombie Vision turns your FlIR ONE Thermal Camera accessory into a realtime zombify viewer. Everyone you see through the camera will look like zombies."
I've only had it since last Thursday.
Looks like it can take video when used with iPhone.
Still, we need control and raw data and I don't think iPhone version will give us that...
Your link has an trailing > in it. Here's the correct link:
http://flir.com/flirone/content/?id=62910
-Phil
Fast time to image (< 0.5 sec)
Looks like it can output all pixel data twice a second.
This one has an SPI interface for the pixel data. That's fine since resolution is so low and update rate is only 9 Hz.
Massimo
I found the Flir One Launch FAQ which has this to say about the original version for the iPhone5,
Those cameras get very very expensive very very quickly. When you are talking about higher thermal resolution, you are really talking about making every facet of your project more complicated and more expensive, for what I personally see is too much data to work with.
Just my opinion.
That is EXACTLY the same thought process that led to me purchasing the one I did, except now I don't much see the need for the more expensive camera (apart from applications where the extra resolution is absolutely necessary). The good cameras at FLIR (and Fluke et al) are over thirty grand.
There has to be a market for it, but the cheapo is more than adequate for probably 80% of what most people really need.
I paid $555.00 for mine on eBay, and if it gets lost or broken I will replace it. EXTRAORDINARILY useful.
The one thing I haven't yet been able to figure out is how the infrared lenses work. A finite resolution, but can you change the lenses for greater field of view? I've been too busy to research the lenses.
Tracy back a few months ago when I was working on my 9v down to 3.3v power circuit I used my iPhone based FLIR to measure the temp of the regulator. It works like a champ.
The other day I had a solder bridge under one of my SMT caps and for the life of me I couldn't find it. Got out the FLIR put it on my iPhone and found the solder bridge.
I love this thing!!
I don't need to connect this to the Propeller for an OEM project or product. Like you and bte2, I just want to look at the heat distribution in Prop (and other) circuits and for other uses around the shop and home. I have access for free to an old iPhone 5 that could use the original Flir One at $150, and I have a recent '6 that is my main phone, that I could use with the new Flir One at $250. I'm fine to use either one of those for the camera screen and software, rather than paying ~$600 for the Flir that includes its own screen. So should it be the $150/iPhone5, or $250/iPhone6? Thus my question to Flir about the claim on their models comparison page that headlines the improvement(?) in resolution. I haven't yet heard back from Flir tech support with a direct answer to my question.
It's kinda annoying that we have to read in frames at 27 frames per second when they only really give 9 frames per second... You have to read in 2 extra copies of each new frame. They say it's for export control.
I saw some Arduino code for Lepton posted, but it's a joke. It only reads and writes to i2c interface, they say Arduino isn't fast enough to handle the video...
This code shows a very basic image (using characters) in Parallax Serial Terminal.
Doesn't do settings over I2C yet, but I'm not even sure I need that yet...
Tracy, yesterday I had another case where I had to find a faulty 0603 cap with my camera. I had several 0603 SMT caps together and the FLIR's resolution was too low to determine exactly which cap was bad. The heat pattern bleeds over on top the other caps, it narrowed it down to 3 or 4 caps but that was it. I then took a small flat blade screwdriver and held it over each cap as I looked at them with the camera. When the blade of the screwdriver was over the hot cap it was obvious on the screen. This is one way to get around the low resolution the camera has. I have only found the resolution to be a problem for very small objects close up.
Another thing that seems to happen often is the visible light image that is overlaid on top of the infrared image get out of sync close up. You'r supposed to be able to drag your finger across the iPhone's screen to align them but it doesn't seem to work well close up. I usually just ignore the overlaid visible light image. It is extreamly useful and for looking at objects 6 or 7 feet away like looking at your homes circuit breaker box. If you get one I just wanted you to know. I would buy another one. It is my go to tool if I'm stuck and can't find the problem with my volt meter.
I found a nice FAQ that compares the new and old FLIR One..
www.flir.com/flirone/.../FLIR-ONE-Android-iOS-FAQ.pdf
Ray, I haven't perused the data sheet at all, but I'm curious if the pixels come out equalized and calibrated in temperature units, or if there are auxiliary calibration arrays? I recall from Phil's thread on the MLX90620 16x4 array that there were several levels of computation necessary to arrive at pixel temperatures.
We already have thermocouple probes for that...
What we want the camera for is to look for uniformity over the area we are heating.
So, absolute temperature is not required, just a sense of relative variations.
There is a "radiometry enabled" mode that I have to figure out though...
Also, I need to turn off frame averaging...
Also, I think the higher resolution lepton might be harder to use with P1...
My code's inner loop runs at 5 MHz and >2 MHz is required.
I'd have to research if there's a way to read SPI data faster than 5 MHz with P1...