ABSTRACT
As discussed in Chapter 1, a colour image constitutes the three mono-chromatic components: R, G, and B, each of which may be considered as a grey-scale image as far as processing is concerned. Mathematical tools for grey-scale images can be applied separately to each of the monochromatic components in order to handle colour images, using the same notation as in Chapter 1, where I = f (x, y) denotes the light intensity function of a pixel defined at the coordinates (x, y). Here f(x, y) is a function in the spatial domain. Methods of image processing in spatial domain contain point operations, local (neighbourhood) operations, and global operations. The result of a point operation is only related to a single pixel. For example, threshold processing for the intensity of a pixel is a point operation. The result of a local operation is related to the neighbouring pixels of a given pixel. In another example, a median filtering has the outcome of a pixel, depending on the intensities of its surrounding neighbouring pixels. A global operation is related to the entire image such as the discrete Fourier transform. A usual neighbourhood includes four or eight neighbouring pixels, as shown in Figure 2.1.
