Analysis of environmental monitoring based on parametric imaging

Leszek Markiewicz a, Jerzy Ihnatowicz b

Division of Biomedical Engineering, Silesian Technical University,
44-100 Gliwice, Akademicka 16, Poland
fax (048)(32) 37-22-25

a/ - lem@biomed.iele.polsl.gliwice.pl
b/ - j_ihn@biomed.iele.polsl.gliwice.pl

Keywords: Environmental monitoring, ecology, database, image processing.

The problem of analysis the results of environmental monitoring can be considered as rather complicated task of comparative image analysis of long series of maps. The problem formulated is as follows: can the coincidence between the fractions of industrial pollution in different years be detected and/or estimated ?

The proposed method based on so-called coincidence matrix will be defined and described.


Figure 1. The principle of calculation of coincidence index

The idea is to decompose the map for two layers: the lower layer consists of pixels below image median value and the upper layer consists of pixels above median value. It should be point out here that this kind of image discrimination will standarize all images. Next the common areas in lower and upper part are determined. The sum of these areas divided by the total map area represents the coincidence.

The set of coincidence values can be considered as the pixels of parametric image. This image can be analyzed separately. Some areas which represent the high-value coincidence between maps can be detected easily. The object specific parameters can be calculated also.


Figure 2. Parametric image of coincidence matrix.

In the paper the basics and some important details regarding the coincidence matrix method is discussed. The result of coincidence analysis between dust falls and its fractions in Upper Silesia region is shown.

Some conclusions on analysis of dust and metal falls is presented.


Figure 3. Several specific features of parametric image and its explanations

A - low coincidence between dust falls in 1970-1981 and Zn and Cd. in 1982-1995
B - low coincidence between Cd. and Cr
C - very high coincidence for cadmium in whole time (1982-1995)
D - low coincidence between Cu in 1982-1995 and Zn and Cd. in 1982-1995
E - very low coincidence between Pb in 1982 and dust falls and its fractions
F - high coincidence between Pb and Cd

The quality of map depends on information about air pollution in region close to Upper Silesia. The analysis of dust falls without information about air pollution in another countries (having a boundary with Upper Silesia Region) is not reliable sufficiently till now. Thus information regarding results of monitoring in Czech Republic and Germany will be necessary for future works.

BENEFIT Special Report Main Page