AUTOMATED IDENTIFICATION AND DETERMINATION OF THE BREED COMPOSITION OF WOOD BY MATERIALS OF DIGITAL MULTISPECTRAL AERIAL SURVEY OF FORESTS

Abstract

Presently, the current task is to automate the determination of the characteristics of forest areas ac-cording to remote sensing data. At the same time there is a large number of automated decryption algo-rithms that are actively used and provide an acceptable result in photointerpretation of medium and low spatial resolution multi-zone images. In processing of high and ultra-high resolution data, both new possibilities for determination of tax characteristics and certain difficulties related to efficiency reduction of algorithms using luminance characteristics arise. The aim of the study is to check the efficiency of the trees automated identification method and to determine their breed affiliation from the images ADS40 of the digital aerial camera. Proposed technique is based on a complex use of methods for controlled classification and identification of objects in the image by areal images. The result of the work done has shown, in the opinion of the authors, a significant efficiency of using the considered methodology on test data and the prospects for its further development.