May 3, 2016:
How does thermal imaging work?
How do you know which type of thermal imaging sensor best suits your requirements?
First, you must understand the three factors that influence thermal image quality:
How much pixel resolution do you need? This depends primarily on your intended use and desired image quality.
• Low resolution: = 160x120 (19,600 pixels)
• Medium resolution: 320x240 (76,800 pixels)
• High resolution: 640x480 (307,200 pixels)
|Most digital camera users never take full advantage of 10 megapixel resolution because they never print their photos on paper large enough for the high resolution to make a difference in print quality.
||Unlike with a digital camera, the full resolution of an infrared camera is always displayed; the highest resolution available is relatively modest by today’s digital camera standards.
Even at 640x480 pixels, a high-definition thermal image only takes up a fraction of today’s computer displays; the resulting print quality will always be optimal.
When choosing a thermal camera, the number of pixels (the resolution) is the most important factor in determining image quality.
How much thermal sensitivity do you need? For many applications, it is important to be able to set a narrow a range to see the smallest temperature gradations possible. Thermal sensitivity varies with object temperature. As object temperature increases, the sensor signal output slope increases. As you view hotter objects, the signal-to-noise-ratio improves. Thermal imagers usually display images in palettes comprised of 256 discrete shades of color or gray. Imagine your target has a temperature range between 0°C and 256°C: Each shade of color or gray would represent 1 degree of temperature variation. If you apply this theory to a scene with a much smaller temperature range, between 25°C and 35°C, for instance, each color would represent just 0.03°C. This is much lower than even the most sensitive uncooled cameras. The result is some display of noise.
As the number of pixels increases, so does sensitivity. Image quality thus becomes increasingly dependent on a process called non-uniformity calibration (NUC). A microbolometer imaging array is essentially an array of tiny resistors; because of their small scale, there are variations in how each pixel responds to an object’s infrared energy. During manufacturing, the infrared camera’s sensor must be calibrated, meaning that the differences in response and DC output for each sensor must be zeroed out.
Increasing the image quality ensures:
√ Greater flexibility to inspect targets at varying distances
√ The ability to see low-thermal-contrast targets
√ More intuitive diagnosis of heat-related problems
√ Improved infrared visible fused image quality due to better matching of infrared and visible camera resolutions
√ The option of incorporating lower-cost and lighter-weight optional lenses
√ More intuitive diagnosis of temperature anomalies