I've been researching CMOS image sensors for a small working-distance, low-light application. From this, I roughly know I need a field of view of approximately 100-120 degrees. However, upon looking at various image sensors, a more common attribute listed is CRA (Chief Ray Angle). I've learned that CRA is a better way of measuring how much of an angle on which light can enter a raw image sensor without a lens. However, FOV is the attribute of choice for measuring how much light can enter a particular lens.
Some image sensor manufacturers shift the microlens on each pixel in order to correct for vignetting effects that occur on edge pixels due to CRA. This can be seen in the below picture:
If this was not done, this would occur:
This would obviously result in lost light, which in turn would produce a vignette effect on the finished image. (Shadows on the edges)
However, I'm confused as to how the FOV would be affected by the CRA value of the image sensor itself.
Overall question:
How does Field of View relate to Chief Ray Angle, and how would this affect choosing a lens for a high FOV application?
Hopefully, I am asking the right questions…
Best Answer
The Field of view is driven only by the size of the image sensor and the focal length of the lens you are using. The CRA angle, as specified by the sensor manufacturer, is dependent on the construction of the sensor. The actual chief ray angle of the lens is determined by the rays passing through the lens, and is dependent on the lens design, not the sensor. Hopefully, the CRA of the sensor and the chief ray angle of the lens in use will match well. If they don't, you will lose some light. They do not have to match perfectly.
You may see the term "telecentric in image space". This means the chief ray angle of the lens is zero on the image side of the lens. (The chief ray hits normal to the sensor plane). This is useful if you want to make the location of an image not move when the image plane is slighty defocused. Having a lens telecentric in image space tends to make it more expensive than non-telecentric, but it is necessary for many metrology applications. A lens that is telecentric in image space would not need the microlenses repositioned as you have drawn here (nice drawing, btw).