An image may be continuous w.r.t x and y co-ordinate and also in amplitude. To convert it to digital form, we have to sample the function in both co-ordinate and amplitude.
- The process of digitizing the co-ordinate values is called Sampling.
- A continuous image f(x, y) is normally approximated by equally spaced samples arranged in the form of a NxM array where each elements of the array is a discrete quantity.
- The sampling rate of digitizer determines the spatial resolution of digitized image.
- Finer the sampling (i.e. increasing M and N), the better the approximation of continuous image function f(x, y).
- The process of digitizing the amplitude values is called Quantization.
- Magnitude of sampled image is expressed as the digital values in Image processing.
- No of quantization levels should be high enough for human perception of the fine details in the image.
- Most digital IP devices uses quantization into k equal intervals.
- If b-bits are used,
No. of quantization levels = k = $2^b$
- 8 bits/pixels are used commonly.