Archive for the ‘General Info’ Category

Now I’ll be getting into a bit more detail on what kind of images a thermal camera produces and explain it to you how and why it is all a matter of interpreting that data and deciding how to show it visually…

How Does a Thermal Image Look Like
As I have already explained in my previous post the sensor of a thermal camera does not operate in the spectrum of visible light, but instead in a part of the infrared light and more specifically the Long-wavelength infrared region (LWI). This is the reason that these cameras do not record in color, though color is still used to make for a better and easier to read representation of the thermal data. Thermal cameras do record only temperature information for each pixel of the image they create instead of color information and then use software to turn that data in something visual and easy to understand and analyze. A special color palette is used with a different color mapped to a different temperature value and as a result you get a strangely looking image where different colors represent a different temperature. One of the most common such color palettes is called Iron and you can see a sample thermal image using this palette in the photo above, though this is not the only one. I’ll get more into the different representations of thermal images in a moment, but first I want to talk about something else as well…

Since thermal cameras do not record visual information, but only thermal one, other the resulting image may not be very detailed and it can even be hard to recognize multiple objects on the image if they have closer surface temperatures. This is especially true if you are using a lower resolution thermal imager like in the case of most affordable consumer oriented thermal cameras. So as a means to make things look better and the lower resolution thermal image to look more like a higher resolution one there is an interesting trick being used to help. Thermal camera manufacturers add a second camera that works in the visual light spectrum, just like the camera on your smartphone, that has a higher resolution than the thermal one and they combine the two images to produce a better looking thermal image as if it was recorded with a higher resolution.

On the image above you can see a monochrome representation on how a thermal image may look like, and next to it how a visual camera of the same thing (my hand in this example) can look like. These two images are recorded by the two different sensors on the same FLIR thermal camera with the visual image actually being double the resolution of the thermal sensor. Do note that even the monochrome representation of the thermal data on the left is just that, a representation of the thermal data in a visual way with false colors (black to white – colder to hotter), the same as when using different colors for different temperatures with other color palettes such as the Iron one mentioned already.

So what happens when we combine the thermal image with the visual one, normally a bit of a mess, though if we take only the contours of objects from the visual image and superimpose them on top of the thermal image we get more of what we need. We are essentially improving the visual detail of the thermal image making it easier to distinguish details and objects on the image without diluting too much the visual representation of the thermal information. Of course there are some other ways to help us combine the data of the two camera sensors in order to make things more useful for the person that will be analyzing the thermal information.

You can also get thermal fusion with opacity or even a picture in picture mode for combining the two different sets of information – visual and thermal. In the first image we are fusing the two images with the thermal one being with some opacity to show the visual image in the background with enough details. The second image just puts the two images in a picture in picture mode with the higher resolution visual image as a background and the lower resolution thermal one on top in the center. It is up to the user to decide which one of these different methods works best in each case and will be more useful for analyzing the thermal information when needed. If you have the original thermal image you can always generate any kind of the available visual representations, however that may not always be the case as you might just be looking at a report with some resulting photos and without the actual thermal data and tools available for you.

False Color Representations of Thermal Images
There is a number of commonly used color palettes that have been named specifically being used for representing the thermal information from a thermal camera in a visual way. You have already seen examples of the most commonly used color palettes – the Iron and the Grayscale ones, but there are some more you can also see being used less often. For example in the movie Predator that I have used as an example in my previous post a variation of the Rainbow color palette has been used to represent different temperatures in the thermal image. Aside from it you may see the Arctic or Lava color palettes used as well and even some custom combinations as it is not that hard to create your own color palette to use.

Here is an example that I have made with a custom color palette that essentially goes from black for colder areas to brighter green for hotter ones… and there you have it a green strange looking hand on an image that represents thermal data. The key thing that you need in being able to analyze the visual information that you see, regardless of the color palette used, is to have a color-temperature chart like the one on the right. Without such a color chart to tie the temperature to a certain color value it is really hard to do a proper analysis of the thermal information. Just imagine using the Grayscale color palette, but reversing it and using black for hotter elements and white for colder instead… if you don’t know that and assume it is a regular Grayscale representation you will be getting confusing results.

Iron color palette

Grayscale color palette

Grayscale REVERSE color palette

Arctic color palette

Lava color palette

Rainbow color palette

Rainbow High Contrast color palette

Here are some videos that actually use thermal imaging cameras for some scenes or even for recording the whole thing to give you a better idea on what kind of images thermal cameras produce. Mind you these videos rely on much higher-end thermal cameras than the affordable consumer options I have covered above, so they are capable of recording high framerate videos with a significantly higher thermal resolution…

Emily Haines & The Soft Skeleton – Our Hell, a music video from 2007 apparently shot completely with a thermal imaging camera. The video uses a grayscale representation for the difference in temperatures with darker objects being colder and lighter ones being hotter.

Coldplay – Midnight, a more recent music video from Coldplay shot almost entirely with a thermal imaging camera, again using the grayscale type of representation of color temperature. They do mix some regular images as well in the video and the thermal range was probably not tweaked so well in order to try and more the colder background from the people that have higher body temperature (like in the first video), though that might be intentional as well.

Earl Sweatshirt – Grief, a music video from 2015 with a similar vision like the first video, shot entirely with a thermal imaging camera and again using the grayscale type of color representation of thermal data. The video does include some really nice and interesting scenes with water and fire.

Route 94 – My Love ft. Jess Glynne, a music video from 2014 that was shot almost entirely with a thermal imaging camera. Unlike all of the three above this one uses different color LUT for the representation of the thermal data, this style is one of the most common ones actually and it is called Iron. The colder objects are darker (black for the coldest), while the hotter ones are lighter (white), but the colors pass through blue, magenta, red, yellow in between the coldest and hottest parts.

Alan Walker – Sing Me To Sleep is a music video from this year and it features some shots made with a thermal imaging camera, though most of the video is shot with a regular camera and some special effects were used in it. Using only some frames and not recording and showing the whole thing in thermal imaging allows you to go for not so high-specs thermal camera. Here the Iron false color representation for the thermal data is used yet again like in the above video.

For most people the first peek at thermal imaging were the scenes from the 1987 movie Predator with Arnold Schwarzenegger where the alien used thermal vision to track his prey. Interestingly enough this was probably the first movie to use an actual thermal imager back in the day when they were still really expensive and far from what they are capable nowadays. This is precisely for a lot of people when talking about thermal cameras the images from the movie pop up in their heads, especially the first one, even though thermal imagers were used in most of the sequels after that, though some apparently resorted to “faking” it with CG as well.

There is a common misconception about thermal imaging cameras and that is regarding what kind of information they actually record with a lot of people thinking that the colorful image they see is actually what a thermal camera records. Well, maybe even that Predator movie is partially responsible for that misconception, but thermal imaging cameras actually do not record any color information. The reason for that is pretty simple – they do not operate like regular cameras that work in the part of the electromagnetic spectrum that is visible (400–700 nm range) and we can distinguish colors in, they operate in the a much higher part of the electromagnetic spectrum (8000–15000 nm).

Thermal cameras record infrared radiation emitted from objects and save it as a temperature information and not color information, this means that each pixel does not have RGB values associated with it that represent certain color, but instead has a temperature value. You can think of this as being more like a grayscale image where the lighter a pixel is, the hotter it normally is, though that is just an example representation. The example with grayscale is also what different color representations of thermal images also do, like in the movie Predator, to make things seem easier to understand a LUT or a lookup table of colors is being used where different temperatures are represented by different colors. I’ll get in more details in a separate post, just want to tell you that the Predator movie for example uses the so called Rainbow lookup table to represent the difference in temperature of the objects in a frame.

Another common misconception about the thermal imagers or cameras is that they are “infrared cameras”, but that is not entirely true and is a confusing term to describe them, even though they actually do work in a part of the infrared spectrum. The term infrared camera is mostly used to describe regular digital cameras or even security cameras with night vision that are designed to be more sensitive in the near infrared region of the electromagnetic spectrum (over 700nm to about 1000nm). Thermal imaging cameras operate in a much higher part of the infrared range or the so called Long-wavelength infrared region of about 8000 to 15000 nm. Regular digital camera sensors are not sensitive enough to even reach close to this region, that is why thermal imaging cameras do rely on special type of sensors and not on regular CCD or CMOS sensors used in consumer cameras. This is also the reason why thermal cameras are much more expensive than a regular digital cameras, though in the last few years they have started becoming more affordable and accessible to regular users and not only to professionals that use them for work.

Affordable Consumer Thermal Imaging Cameras
Here is a list of some more affordable thermal imaging cameras that you can purchase and play with without having to pay too much to get your hands on the technology. Of course there are some limitations that you can expect such as the lower resolution of the thermal imager and the low framerate you will get if the camera is capable of recording videos. The recent spike in interest in thermal cameras was pretty much caused by the availability of inexpensive sensors such as FLIR Lepton for example. These small and inexpensive sensors quickly found their way into accessories for mobile phones that add thermal vision capabilities to your device. The fact that the phone takes the role of processing and display device allows to greatly reduce the extra cost of these thermal imaging devices as compared to traditional all-in-one solutions.

FLIR One Thermal Imager for iOS and Android Devices
I have been using standalone thermal imaging cameras in my work for a few years already and when the first FLIR One came out I was really curious how well it would work, so I did not wait much to get one. Back in the time it was pretty much the first thermal imaging smartphone accessory that relied on the first generation of FLIR Lepton sensor to come out on the market (80×60 pixels thermal resolution). The product was initially designed only for Apple’s iPhone 5/5S as it came in a special accessory case that attached only to these iOS devices.

There were a couple of interesting features available that were introduced with it that were new to the not so high-end thermal imaging cameras that were available back then. These included the ability to actually record thermal videos with it (low resolution and framerate), but still something that is not yet available to the more affordable standalone thermal cameras. This was possible thanks to using the pretty fast processing power of the iPhone to work with the data captured with the sensor. The other one is the ability to record two images and superimpose them in order to create a higher-resolution looking end result, the lower resolution thermal image gets upscaled and on top of it a higher resolution visual image is overlaid providing contour of actual objects. The end result of the so called FLIR MSX blending gives the impression of actually using a higher resolution thermal imager than what you actually have in terms of thermal imager resolution.

The second generation FLIR One Thermal Imager that was introduced later on was made to be compatible with a wider array of devices and not be fixed only for a specific brand and model(s) of smartphones like the first gen. It became available in a more compact form and available in separate versions for both iOS and Android devices. It also offered some slight improvement in the specifications such as higher temperature sensitivity and apparently a better thermal resolution, not to mention the much more convenient and smaller size of the whole thing that you can easily carry in your pocket for example.

For more details about the FLIR ONE Smartphone accessory…

Seek Thermal Compact Imager for iOS and Android Devices
The other main player in the smartphone accessory thermal imaging cameras was a company called Seek Thermal introducing their Seek Thermal Compact smartphone accessory to rival the FLIR product. Back when they did this their product was more advanced in terms of specifications and came in a more compact fork with different versions for iOS and Android OS available, but still had some trouble getting popular initially. The reason for that is in the fact that FLIR is one of the specialists and the most talked about name in the thermal imaging cameras for professional use, so a tough competitor. There were however some other issues that Seek thermal had, such as the initial availability of their products only for the US market and the not so good performance of their first product even it being with better specs. The last reason was one of the most important as while FLIR has many years of experience and just used it in their consumer product in terms of hardware and software as well, Seek Thermal was really new to the whole thing and needed some time to polish things up.

At the moment Seek Thermal is doing much better with multiple product offerings with different specifications and for different user needs, including a more professional solution with even higher resolution thermal imaging sensor, manually focusable lens, higher refresh rate and so on. The company has also managed to polish their software as well in the meantime, so that it is more usable and provides better results than initially had. So if you are interested in thermal imaging cameras and being able to add such functionality to your smartphone you might want to also check Seek Thermal as well and not only what FLIR offers.

For more details about the Seek Thermal Compact Smartphone accessory…

Other Affordable Thermal Imaging Cameras
There are of course some other affordable entry level products in the form of thermal cameras that you might be interested in such as the standalone FLIR TG165 Spot Thermal Camera that is also based on the FLIR Lepton sensor. There is also the standalone solution from Seek Thermal in the form of their Reveal, RevealXR and RevealPro product line. It is not only these two companies however, there are already some other quite interesting alternative products. For example there is the CAT S60 smartphone with an integrated thermal imaging camera (FLIR Lepton), so that you don’t need to get a separate accessory for that functionality. There are also some other interesting projects for people into DIY such as the DIY-Thermocam project, so you might want to check that one out as well…