What’s the MTF for image sensing?
A camera is one of those things that we use on an everyday basis, and it is also an engineering marvel once you start digging deeper. The camera in our phones works at the intersection of CMOS chips and optics. A lens focuses an image onto a CMOS sensor, which then returns the intensity of light at each pixel. Couple that with filters that can be used to figure out the intensity of colours, and you have a modern colour camera.
An important characteristic in image formation is the modulation transfer function (MTF), which describes the frequencies at which a component can create a defined object. As an example, if an image with perfect contrast goes through a lens or image sensor, the image loses information and the lens or sensor is not able to completely transfer all the information:
Taking a discrete example will help us understand why. Let’s say there is an image sensor with a pixel pitch of 5μm (meaning the pixels are 5μm big). If you apply a sine wave signal to the sensor with a period of 10μm, it will have maximum contrast as all information is kept within each pixel (top half of the image below). But if you had a sine wave signal with a period of 5μm, it results in a grey image because each discrete pixel takes an average of the signal it captures, and since the average of the sine wave is 0, the sensor cannot preserve any information (bottom half of the image below):
Understanding this concept is important even outside the context of cameras, as it also applies to areas where images are being transferred from one place to another using optics, such as lithography and chip fabrication. Characterizing the MTF during lithography setup is crucial in ensuring the desired patterns are transferred to the chip in your device so you’re able to read this article!
