Polishchuk B. Processing methods of high resolution satellite images for classification of forests (on the example of Pre-Carpathian region)

This paper consideres the theoretical framework and the possibilities of using space information for solving the basic problems of forestry and effective management of Ukraine forestry and the Pre-Carpathian region in particular. Basic forestry tasks solved using remote sensing aerospace systems of various resolution and spectral range are systematized. A method for determining the modulation transfer function of remote sensing aerospace systems based on the normal law of distribution of various factors impact on the image is processed. The number of required measurements to obtain the values of blur are grounded. Existing mathematical models of object classification are considered and a model of hybrid classification of forests is proposed. The main stages of this model are the following: determining of the optimal number of classes based on the results of uncontrolled classification, creation and analysis of educational standards and implementation of controlled classification. To assess the effectiveness of the proposed method, testifying information is obtained from field studies and measurements with fixing forest taxation parameters of selected polygons and coordinates of boundary points are determined. The area of cutting is determined by the pictures taken in different periods.