Texture Classification in Bioindicator Images Processing

The section deals with classification of microscope images of Picea Abies stomas. There is an assumption that a stoma character strongly depends on the level of air pollution, so that stoma can stand for an important environmental bioindicator. According to the level of stoma incrustation it is possible to distinguish several classes of stoma structures. A proposal of an algorithm enabling the automatic recognition of a stoma incrustation level is a main goal of this study. There are two principles discussed in the paper: The first principle is based on gradient methods while the second one uses a wavelet transform. Possibilities of application of mentioned attitudes were investigated and the classification criteria distinguishing the stoma character were suggested, as well. The resulting algorithm was verified for a set of four hundred real images and results achieved were compared with an expert’s sensual classification. Selected methods of image preprocessing as noise reduction, brightness correction and resampling are studied, as well.