The infrared thermal imaging technology was once only used for high-end applications such as national defense and military, but in the past few decades, it has gradually become a more mainstream technology. With the growing popularity of handheld thermal imaging cameras in maintenance and electrical troubleshooting applications, there is a growing awareness of the technology's industrial benefits.
In 2020, we encountered the new crown epidemic again, and a large number of infrared thermal imaging solutions flooded into the market logically. Various thermal imaging sensor technologies and cameras are used for heat detection, although technically thermal imaging sensors can only measure skin surface temperature.
Nonetheless, infrared thermal imaging remains mysterious to many end users. Even the most skilled personnel may be at a loss for non-visible light imaging techniques. This is not unusual, as humans lack the visual ability to visually perceive temperature.
To better understand the performance of infrared cameras and thermal imaging cameras
, users must understand how infrared thermal imaging cameras work and the physics involved. Unlike standard machine vision cameras that operate in the visible spectrum (400nm to 700nm band), infrared camera and thermal imaging technology covers a wider spectral range, which is subdivided into three main bands: 0.9μm to 1.7μm The band belongs to short-wave infrared (SWIR), the 3μm to 5μm band belongs to mid-wave infrared (MWIR), and the 8μm to 14μm band belongs to long-wave infrared (LWIR).
Spectral bands are primarily defined by the characteristics of the detector technology in various types of cameras. The spectral bands come from the sensitive wavelengths of the detector material. Depending on scientific principles, the physics literature may classify infrared spectroscopy in different ways.
A brief introduction to long wave infrared
LWIR collects light in the 8μm to 14μm spectral band, the wavelength range of the most available thermal imaging cameras. In fact, according to Planck's law, ground targets are primarily emitted in LWIR. LWIR system applications include thermal imaging/temperature control, predictive maintenance, gas leak detection, scene imaging across a very wide temperature range (and requiring a wide dynamic range), smoke imaging, etc... the two most commonly used The uncooled detectors in LWIR are amorphous silicon (a-Si) and vanadium oxide (VOx), while the cooled detectors in this region are mainly HgCdTe.
Microbolometers: a more economical thermal imaging technique
The real thermal effect at room temperature and below is manifested in the 3μm and above bands. Imaging devices capable of capturing these thermal effects are often considered true thermal cameras. The term "infrared camera" doesn't just refer to this part of thermal imaging devices - since most of the signals they capture come from the long-wave infrared radiation.
MWIR detectors can also be used in thermal imaging. However, they have a common disadvantage, which is that they are very expensive. The median selling price for a 640 x 512-pixel detector is about $70,000. These detectors are expensive because they must be cooled to about 75K (or -198.15℃). The detector material itself is very sensitive to thermal radiation, thus causing the sensor to saturate immediately at room temperature.
In modern MWIR cameras, cryogenic cooling is achieved by a closed-circuit Stirling cooler located inside the camera body. In the past, cooling of such cameras required the use of large cylinders filled with liquid nitrogen.
A more economical option is a thermal imager with an integrated microbolometer detector. Depending on pixel resolution, detector noise levels, and temperature measurement accuracy, these cameras can start at less than $1,000 with a resolution of 80 x 60 pixels. Microbolometers work quite differently from typical photon-capturing detectors and are primarily based on tiny thermally resistive pixels. Some of these cameras mainly use thermoelectric cooling elements, which are easier to operate. When these pixels are exposed to infrared radiation (heat), their resistance changes. No low-temperature refrigeration is required, the operation is simpler and the cost is lower.
Each pixel in an LWIR camera has a physical mass that needs to capture thermal radiation from the object it is pointed at to heat it. This gives a fixed time constant for the time it takes for each pixel to warm up before the camera reads the resistance change. This constant is usually between 8 and 14 milliseconds, depending on the pixel size. The downside of such detectors is that the time constant presents challenges when it comes to imaging moving objects.
Eight milliseconds may seem like a short time, however, depending on the camera's field of view and the speed of the imaged object, there may be noticeable motion blur in the captured image. During the integration time (i.e. the time constant), motion blur occurs when part of the object passes the detector pixel. In other words, the pixel may not have fully integrated the thermal radiation it is trying to capture before the object moves to an adjacent pixel. As a result, this can cause temperature averaging effects, which can lead to measurement errors and other problems.
Motion blur is not the only type of blur in thermal imaging. Because the contrast in thermal images is caused by temperature changes, most thermal images appear blurry. This blurring is not the result of focus or lack of focus. More precisely, this is caused by physical thermodynamic functions.
Heat energy flows from warmer regions with higher energy to cooler regions with lower energy. This behavior is completely dynamic, resulting in temperature transitions or thermal gradients. Temperature changes are represented in thermal images as changes in brightness: white represents hotter areas, black represents cooler areas, and a gray transition occurs between warmer and cooler areas.
These transitions make the edges of the image look blurry. This effect is not typically seen in standard machine vision applications, which rely more on the effect of light reflecting off a surface or feature. This reflection pattern is constant, and so is the contrast it produces in the image. Thermal images only appear sharper when radiance changes, or when warmer areas are thermally isolated from surrounding areas. It is this dynamic behavior caused by thermal diffusion that suggests that thermal imaging may have more to do with signal processing than image processing.
Emissivity is probably the most important phenomenon to understand when studying thermal imaging cameras. So it tends to be one of the hottest topics in thermal imaging courses and seminars. Simply put, emissivity characterizes the ability of a solid to radiate infrared energy. Emissivity is mainly composed of three components: reflection, transmission, and radiant energy. The sum of these factors must equal 1.
Since most materials do not transmit infrared radiation, imaging is primarily about reflected and radiated energy. In this case, the derivation process can make it difficult to measure the temperature of the heat-reflecting object. For example, trying to tell the temperature of a gleaming stainless steel tank is considered an impossible thermal imaging application unless the emissivity of the tank surface can be changed. If permitted, paint black paint can be applied to an area of the tank to increase its emissivity to 0.9 or more. Using thermal conductivity, this high emissivity coating will absorb the temperature of the tank surface. The coating then helps transmit energy to the thermal imager, enabling accurate temperature measurements.
When it comes to applications involving low emissivity surfaces that cannot be altered, then measurement by contact methods (eg connecting a physical thermocouple) may be required.
Another factor to consider when using a thermal imager in machine vision is the available spatial resolution of the thermal imager. For commercial applications, thermal imagers have a maximum resolution of about 1.3 MP, with more economical cameras offering 640×480 or 640×512 pixels. This resolution pales in comparison to state-of-the-art machine vision cameras, which offer 70 MP or even 100 MP. Therefore, infrared cameras still have a lot of room for improvement.
Lens materials for thermal imaging cameras are special. The most typical one is germanium (Ge). Standard borosilicate glass blocks mid-wave infrared and long-wave infrared light, making it unsuitable as an optical material for thermal imaging cameras.
Camera manufacturers have to calibrate their lenses according to the camera itself, so many camera manufacturers are also their lens suppliers. Therefore, it is not uncommon for each thermal imager to offer only 1-5 lens options, which complicates the design of the imaging system.
The situation is further complicated if the thermal imager also needs an enclosure to protect it from harsh environments. In this case, the viewing window must also be equipped with infrared transmissive glass made of germanium or other suitable material.
Despite these challenges and shortcomings, thermal imaging cameras are becoming increasingly important in industrial and non-industrial imaging applications. Thermal imaging cameras are sure to shine in their own unique way.
Quanhom is a professional custom infrared lens manufacturer
. Our one-stop solutions to complex challenges in the defense, security, and commercial applications are recognized by customers worldwide. A team of experts continuously develops first-class thermal infrared technology. Seasoned engineers have decades of experience designing complex infrared optics-related products. Quanhom's talented team has created many success stories for different applications such as thermal imaging sights for outdoor and defense use, thermal imaging monoculars/binoculars, border, and coastal security, maritime applications, and UAV infrared payloads.