Pixel Pitch in Infrared Thermography

Among the key parameters that define the performance of a fixed-mount infrared camera is pixel Pixel pitch is the distance between the centers of adjacent pixels on the thermal sensor. This feature plays a crucial role in both image quality and thermal measurement accuracy.

This article highlights the technical significance of pixel pitch, discusses current trends in sensor design, and compares two infrared camera models from Optris: the PI640i with a 17 µm pixel pitch and the Xi640 with a 12 µm pixel pitch.

Pixel pitch, measured in micrometers, refers to the space between individual detector elements on an infrared sensor. Smaller pixel pitch values allow for more pixels to fit in a given sensor area, leading to better geometric resolution. However, in thermal imaging, pixel pitch also impacts thermal sensitivity. Larger pixels capture more infrared radiation over time, resulting in stronger signal strength and a better signal-to-noise ratio.

Recent advancements in microbolometer technology have made it possible to produce thermal sensors with reduced pixel pitch. Arrays with 12 µm and even 10 µm detectors are becoming more common. This move towards smaller pixels is driven by the need for more compact and lightweight camera designs, which are easier to integrate into smaller or space-limited industrial systems. Additionally, smaller pixels allow for cheaper production by optimizing the use of sensor material and enabling higher pixel counts in a given sensor area. This leads to detailed thermal imaging in smaller optical formats.

Despite the benefits of smaller pixel pitch, larger pixels are still preferred in applications where accuracy in temperature measurement and thermal sensitivity are essential. The PI640, which has a 17 µm pixel pitch, illustrates this point by providing strong and reliable thermal performance. The larger pixel size allows for greater infrared energy absorption per pixel, which leads to stronger thermal signals and better signal integrity. This characteristic results in a lower Noise Equivalent Temperature Difference (NETD), which improves both the accuracy and repeatability of measurements. Moreover, larger pixels are less vulnerable to thermal noise, especially in low-emissivity or low-temperature situations, and they offer greater durability in tough industrial environments.

These features make sensors with larger pixel pitches ideal for continuous condition monitoring, metal processing, electronics manufacturing, and other precise thermal measurement tasks.

The following images show the difference in thermal image quality between the PI640 and Xi640, both capturing the same architectural structure.

The reason for this becomes clearer when considering the optical behavior at the detector level, as shown in Figure 4. This figure presents a top view of bolometer arrays with different pixel pitches. In this example, the optical system is the same for both configurations. Under these conditions, the imaging system’s performance is limited by the number of pixels and the basic physics of light propagation, particularly the diffraction limit set by the optical aperture.

At the center of this limitation is the Airy disk, a diffraction pattern that indicates the smallest possible point of focus a lens can create. When light passes through a circular aperture, like a lens, it does not come together at a perfect point but creates a central bright area surrounded by concentric rings. This central region is the Airy disk, and its diameter defines the diffraction-limited spot size of the optical system. The diameter of the Airy disk is inversely related to the aperture size and directly related to the wavelength of light. Longer wavelengths, like infrared, naturally create larger Airy disks compared to visible light.

In cases of reduced pixel pitch, such as moving from 17 µm to 12 µm, the sensor has a higher pixel count, but the optical resolution doesn’t change because the Airy disk size does not decrease. Thus, the same amount of infrared radiation spreads over more, smaller pixels. This means multiple pixels must be sampled and combined to capture the total radiant power from a single diffraction-limited spot. In practice, this can increase the impact of readout noise and lower the signal-to-noise ratio, which negatively affects temperature measurement performance.

Moreover, theoretical improvements in resolution that come with smaller pixels often ignore real-world limits in lens design and manufacturing. Making optics that can fully resolve features at the scale of smaller pixels without introducing distortions or alignment problems is complicated and expensive. Flaws in lens production, coating uniformity, and mechanical assembly tolerances often restrict the effective resolution and contrast that the system can achieve, regardless of the sensor’s pixel pitch.

In this experiment, we compare two infrared cameras with different pixel pitches, one with 17 µm and the other with 12 µm, to see how well they measure temperature.

The analysis looks at encircled energy, which shows how much infrared radiation is gathered in a specific area of the image. To assess this, we first light up the entire image plane using a large-area blackbody source. This guarantees that every pixel gets full infrared intensity. We project the image much larger than a single pixel to avoid undersampling.

An adjustable aperture is placed between the blackbody and the camera to control how large the visible radiating object appears. When the aperture is fully open, the central pixel captures the most energy. As the aperture closes, less energy reaches the camera’s focal plane, and diffraction spreads the energy across multiple pixels, reducing the amount focused on any single pixel.

Figure 5 illustrates the experimental setup: a blackbody at 100 °C with a variable aperture in front. The cameras observe the setup, and we analyze the temperature response across a single line of pixels in the focal plane array for simplicity.

The thermal camera with the 12 µm pixel pitch shows the greatest deviation in temperature measurement. It can pick up finer details due to its higher spatial resolution, but the actual minimum measurable feature size is still restricted by optical effects and not just pixel resolution. In infrared imaging, diffraction and the optical system’s properties limit how sharply the energy can be focused onto a pixel.

These results show that even with similar optics and field of view, the temperature readings of a small object can differ greatly between cameras. Both cameras have almost the same IFOV values at 1 meter (about 1 mm), but their temperature measurements vary due to how energy is spread and captured at the pixel level.

In this case, the camera with a 17 µm pixel pitch does a better job of measuring temperature accurately. Even though it has larger pixels, it captures more infrared energy per pixel. This leads to stronger signal strength and improved temperature accuracy. To measure a small object in this situation, the 17 µm pitch camera needs to cover at least a 3×3 pixel area, while the 12 µm pitch camera requires a 4×4 pixel area to capture enough energy and achieve reliable temperature readings.

Table. 1. Measurement results of a camera with different pixel pitch.

While smaller pixels, such as 12 µm in the Xi640, provide higher spatial resolution and more compact camera designs, they do not always improve temperature measurement accuracy. In fact, larger pixels, like the 17 µm pitch in the PI640i, often perform better in thermal applications, especially when precise, non-contact temperature data is needed.

Larger pixels capture more energy, which improves the signal-to-noise ratio and reduces measurement errors. Ultimately, the resolution limit in thermal imaging depends not only on pixel pitch but also on the quality of the optical system. In most industrial applications, the advantages of higher thermal sensitivity from larger pixels outweigh the theoretical gains in resolution from denser sensor arrays.

This finding is important for applications like early fire detection systems, where quick and precise identification of small hot spots is essential. Infrared cameras with smaller pixel pitches may offer higher spatial resolution, but they can struggle to detect tiny targets at the right temperature in time. This happens because each pixel captures less infrared energy, leading to weaker signals and possibly delayed or inaccurate temperature readings. If the thermal signature of an early-stage fire is spread over several small pixels, diffraction and optical limits can cause the peak temperature to seem lower than it really is. On the other hand, cameras with larger pixel pitches absorb more energy per pixel, allowing for more reliable detection of small, intense heat sources. In safety applications where timing is critical, like monitoring vital infrastructure or industrial areas for early fire risk, the ability to obtain accurate temperature readings—even from small or distant sources—can significantly affect response time and system effectiveness.