How Accurate Are Thermal Imaging Cameras?

Infrared cameras detect the heat signatures of different objects and measure their temperatures. However, the ability of thermal cameras to show accurate temperature readings by minimizing the margin of error largely depends on proper calibration and several other factors.

There’s no point in purchasing an expensive thermal camera if you don’t know how accurate are thermal imaging cameras in detecting and showing temperature readings.

To understand how the accuracy specifications and sensitivity of thermal cameras vary, it’s essential to know the basic working principles of these cameras. So in today’s article, we’ll discuss how thermal imaging cameras work while highlighting the factors impacting their reading accuracy.

What Is A Thermal Camera?

Table of Contents

A thermographic image of a person's hand showing different temperatures in different colors, from blue indicating cold to red indicating hot, this may indicate inflammation of the joints. Red palm. Collage. What Is A Thermal Camera.

A thermal imaging camera is an electronic device with a fully-integrated visual display specially designed to detect heat energy and take temperature measurements of various objects.

The main component of a thermal camera is its heat sensor, which is primarily responsible for measuring the surface temperature and infrared radiation of any object. This heat sensor works alongside thermal imaging systems and traditional image-capturing technologies, allowing you to visualize and quickly detect overheated blackbodies.

Since visible light is the only part of the entire electromagnetic spectrum that you can see, ordinary cameras can’t capture infrared radiation. Your only solution is to use thermal imaging technology if you want to detect the heat signatures of an object. Once you point a thermal imaging camera at any object, its sensor will allow you to take accurate temperature measurements and view the invisible infrared radiation.

Infrared wavelengths exist between visible light and microwave spectra. In a modern infrared (IR) camera, you can see the heat radiation from different objects in different colors, known as the color map. However, grayscale displays are preferable in specific applications because they are more effective in capturing fine details and reducing visual busyness.

A thermal imager with a colored thermographic display shows components with high body temperature and heat energy in yellow, orange, and red shades. On the other hand, cooler components or objects usually appear in blue and purple colors, while green primarily represents room temperature.

Since thermal cameras help in temperature measurement by tracking heat signatures, they don’t require visible light and can identify heat sources even in dark and obscure environments. So, if you are wondering which thermal imaging is the best, look for models adept at detecting temperature differences to create authentic measurement data.

How Do Modern Infrared Cameras Work?

Before purchasing a thermal camera, you might ask – how does a thermal camera work? An infrared or thermal imaging camera works by detecting and measuring the heat signature of an object or area. The camera is fitted with a telephoto lens to allow IR wavelengths to pass and focus them on a heat or IR sensor array, which detects and reads the radiation.

The sensor is essentially a grid of pixels, and every single pixel reacts to the IR wavelengths and converts them into electronic signals. These signals are then sent to the processor inside the camera to develop color maps of various temperatures using a range of algorithms. This color map is projected on the display screen to view the measured temperature.

Most thermal imaging cameras also come with a shooting mode like normal digital cameras that work in the visible spectrum. Using these cameras, you can compare the two shots – one taken in the IR mode and the other captured in the normal mode. Hence, you can identify the specific problem areas by comparing the IR mode with the normal mode.

Thanks to their heat sensitivity, thermal imaging cameras are also used in different industries, such as building construction, law enforcement, medicine and healthcare, and plumbing and electrical maintenance. Furthermore, since they can measure temperature and detect heat signatures, thermal imaging systems are widely used for object detection in vehicles under unfavorable driving conditions, such as fog, storms, rainfall, and absolute darkness.

Do Thermal Imaging Cameras Work Better At Night?

Even though a thermal camera usually works better at night, it has nothing to do with the availability of visible light. The ambient and core temperatures of unseated objects or components are typically lower at night due to the absence of sunlight. This makes it easier for any thermal camera to display warmer areas or objects at higher contrast than the surrounding.

All earthly components, including roads, vegetation, buildings, construction materials, and metal structures, always absorb heat energy from the sun, even on a relatively cool day. The more heat these objects absorb throughout the day, the more their core temperature rises, making them warmer.

Under such circumstances, it becomes harder for a thermal imaging camera to distinguish these objects from any other warmer component the camera tries to detect. In contrast, thermal imaging cameras will show a distinct contrast in the temperature reading during the night. The cameras will show a sharper temperature contrast if the warmer objects are shot in the middle of the night compared to just after sunset.

Even if you’re shooting in full daylight, you’ll get more accurate readings and better image quality in the early morning than in the afternoon. It’s simply because the objects don’t start absorbing heat from the direct sunlight during the early morning and maintain a cooler core temperature.

Can A Thermal Camera Work Through Glass?

Unlike traditional digital cameras, thermal imaging cameras can’t work through glass. Even though a regular glass sheet allows visible light to pass through, it acts like a mirror for infrared light and can’t create a thermal image. Thus, IR camera lenses are made of zinc selenide or germanium instead of glass.

If you point a thermal imaging camera toward a glass window, you won’t see a clear thermal image of what’s on the opposite side of the window. Instead, you’ll get a blurry image, which will most likely show a vague and hazy reflection of you standing there holding the camera.

However, some infrared frequencies can pass through a standard glass sheet. Also, some special types of glass allow infrared wavelengths to pass through to some degree. For instance, tinted car windows usually offer better measurement accuracy compared to any standard household glass glazing, which you usually see around buildings.

That said, in most cases, you’ll get a largely obscured thermal image because of all the infrared reflection from the glass displayed in different degrees of opacity. It’s safe to say that the object you want to view using the thermal imaging camera will lack temperature contrast and significant details. So, you can’t use an infrared camera to get accurate temperature readings through glass or any other high-reflective surfaces.

Can You See Through Walls Using A Thermal Camera?

Technically, you can’t see through anything using a thermal imaging camera. Rather, it detects and measures the surface temperature of any object in its field of sight. If you point a thermal camera at a solid surface, like a wall, it will detect the heat energy radiating from the wall surface.

Since most building materials are designed and insulated to absorb heat, thermal imaging doesn’t usually reveal any significant image regarding what’s going inside a building wall or vice versa. However, an IR camera can register extreme heat radiating from inside a wall because the wall gets overly heated up. That’s why thermal imaging cameras effectively detect house fires even from a safe distance.

Similarly, a thermal imaging camera is sensitive enough for accurate body temperature measurement if a person stands close to the opposite side of a relatively cold and thin wall. But the camera can only register the temperature contrast if the person stays in their position long enough for their body temperature to be partially transferred through the wall materials.

Do Thermal Cameras Work Underwater?

Thermal cameras don’t work very well underwater because of water’s high reflectivity, which blocks most of the infrared wavelengths from reaching the camera lens. Just like opaque objects block visible light wavelengths, and we can’t see them through a solid surface, IR sensors can’t detect the infrared frequencies through the water. So, the IR waves a thermal camera usually detects are blocked by water.

Another problem with using thermal cameras underwater is related to specific heat and thermal conductivity. The specific heat of water is higher than air, making its thermal conductivity almost 25x higher. Thus, it’ll take almost four times the energy to increase or decrease the temperature of a component by a degree underwater compared to open air.

In simpler terms, an object will gain or lose its own heat energy much faster than water. Therefore, it’s naturally much more difficult for thermal cameras to detect temperature contrasts if the objects are submerged in water than in the open air.

Camera Accuracy Specifications

Most thermal cameras have a data sheet showing accuracy specifications, like 2% or ±2℃ of the temperature reading. This accuracy specification is calculated using a measurement uncertainty analysis technique known as RSS or Root Sum of Squares.

It allows you to calculate partial errors for every variable of the equation used for temperature measurement. Then, you can determine the square values of every error term, add them, and evaluate the square root. Even though the equation might sound a bit complex, it’s pretty straightforward. But calculating the partial errors can be difficult.

These partial errors can occur due to several variables while measuring temperature using a thermal camera. The variables can be as follows:

Emissivity
Reflected ambient temperature
Transmittance
Atmospheric temperature
Camera response
Temperature accuracy of calibrated blackbody

Uncertainty Equation

Once you’ve determined reasonable values for partial errors for each of the terms mentioned above, you can calculate the overall uncertainty equation for the temperature measurements. The uncertainty equation will be:

Total Error = (T1)2+(T2)2+(T3)2…..etc.

In this equation, ΔT1, ΔT2, and ΔT3 signify the partial errors of each variable.

Sometimes random errors add up to take your temperature readings far away from the actual value, while at times, they might work in opposite directions, canceling out each other. Calculating the RSS helps you to get the most appropriate value to determine the overall error specification. All IR camera data sheets follow this method to show their error specification.

All the calculations discussed here will be only valid if you’re using the thermal camera inside a lab or at a short distance of fewer than 20 meters outside. Using the camera for longer distances will introduce more uncertainty in the readings due to atmospheric absorption and emission.

While performing an RSS analysis for any modern thermal cameras under standard laboratory conditions, the final value is kept around 2% or ±2℃. It’s a reasonable accuracy rating for all camera specifications.

Ambient Temperature Compensation

Ambient temperature compensation is a critical step in calibrating a thermal camera. Any infrared camera responds to the total infrared wavelengths that hit the camera’s sensor. For a well-designed camera, most of this radiated energy will be generated from the black body, and very little radiation will be generated from the thermal camera itself.

However, you can’t totally eliminate the radiation detected from the components surrounding the optical path and detector of the cameras. So, without proper ambient temperature compensation, any changes in the temperatures of the lenses or camera body will greatly affect the readings of the camera.

The easiest and most effective way to achieve compensation is to take the temperature measurements of the thermal camera and its optical path at three different locations. Then, you can calibrate this measured temperature data while calculating the calibration equation. It will ensure that you get correct readings across the whole operating temperature range of -15℃ to 50℃.

Ambient temperature compensation is even more critical for thermal cameras that face temperature fluctuations or are used outside. But even after ambient temperature compensation, make sure that the camera fully warms up before you can take critical measurements.

Always keep the camera and the optics away from sunlight or heat sources because any changes in the temperature of the camera and optics will adversely affect the measurement uncertainty.

Other Considerations For Measurement And Accuracy

1. Emissivity

Emissivity is the ability of an object to emit infrared energy. You need to calculate the emissivity of your target object and enter the information in the camera for it to show correct readings. If the subject is highly reflective, you can coat its surface with a non-reflective paint before taking the readings.

All thermal cameras allow you to determine the correct emissivity of different subjects. Even if you make any mistake, you can use FLIR R&D software to rectify the emissivity during post-analysis.

2. Spot Size

Spot size defines the area of your target each camera pixel covers. Say a thermal camera with a 25° lens measures the temperature of a lit match placed at 60 feet. Each pixel usually covers a square inch of the entire scene.

However, the match is only ⅛ square inch, much smaller than what a pixel covers. Most of the infrared energy comes from the subject’s surroundings, while only 1/64th of the measurement comes from the match. Hence, the camera will naturally show inaccurate temperature readings of the lit match.

You can use a telescopic optic or move closer to the subject, so the ratio between the pixel size and the lit match is much closer to a 1:1 ratio. A 10 x 10-pixel grid coverage of the smallest target object usually offers the closest absolute temperature accuracy. However, you can get pretty good readings with a 3 x 3 or a single-pixel grid spot size.