ABSTRACT
Digital image filtering has emerged from the theory of signal processing. It consists of a group of methods that analyse the signal in spatial domain or frequency domain. A digital image is hence represented by a discrete and finite zone of discrete amplitude. It constitutes spatial representation of an object of a two-dimensional or three-dimensional scene or of another image. Most systems operate in square meshes. It is possible to simulate a hexagonal grid on a square grid, however, by means of processing. The most common digital representation used for images is the matrix representation. Each image pixel is associated with a cell of the matrix. Image processing is the analysis of two-dimensional digital signals. The definitions given in the preceding section are extendable to two-dimensional signals such as images. The coefficient of normalisation is generally defined as a function of the weightages of the convolution filter.
