Image Fundamentals

  • (Some) Light reflects off the scene, and hits the optical lens which focuses the light
  • Passes colour filter
  • Light enters the imaging sensor

We’ll go into depth on each of these points now.

Imaging Sensor

Sampling

Sampling is the discretization of space (since the real world is continuous, but the digital is discrete).

Spatial Resolution

Grid spacing size determines spatial resolution (how detailed the digital image is)
- denser grids more pixels, more detail

Sampling in 1D

Quantization

The discretization of intensity values.

Quantization is a more general concept which means to go from a larger (usually continuous) set to a smaller finite set i.e. it just means discretization.

Colour Filter Array (CFA)

Filters for the three primary colours.

Bayer Patterns

The CFA is composed of 3 filters which form the RGB mosaic.

![Info]- Why double the green?
The green filter has double the number of cells, because the human eye has double the number of green absorption cells.

So it is to mimic human physiology.

CFA Interpolation

We then have to recover the complete RGB image from the Bayer Pattern array; we do this through interpolation.

Linear Interpolation

Refer to Data Analytics

Bilinear Interpolation

Given 4 pixels, we can interpolate between two pixels (to get ) and then between the other two pixels (to get ).

Gamma Correction

Grayscale Image

Lecture Question 1: What is the storage for a 128x128 grayscale image which takes 16 different intensity levels?
Answer: bits.
Lecture Question 2: Amount of storage for an RGB image of the same dimension?