Development of a Scientific Methodology for Forest Resources Assessment Using Automated Interpretation of Multispectral Remote Sensing Data

The project initiator was the National Academy of Sciences of Ukraine which envisioned the need to implement new methods of emergency monitoring for human and environmental protection, provided for in the Law of Ukraine on the Protection of the Population and the Territories from Emergency Situations of Human and Natural Origin.

The aim of the project was to develop a scientific method of forest resource assessment using an automated interpretation of multispectral remote sensing data and to prepare recommendations for the monitoring of forest resources of Ukraine.

The project implementers analyzed the potential of developing an algorithm, based on input data in the form of high-resolution (30-m and better) multispectral satellite images for reliable decoding and thematic interpretation of comprehensive sets of remote sensing data with a minimal use of in-situ information. A method of classifying vegetation groups using ERDAS Imagine was studied.

Among the achievements of the project were a validation of theoretical approaches, development of a classification algorithm for objects observed with multispectral imagery, including a preliminary assessment of the information content of the data, and implementation of a data merge approach. Software was developed for training of test sites, identifying information content of the channels of multispectral satellite images, as well as classifying and assessing classification precision.

Forest resources conditions can be assessed successfully with the application of an automated interpretation of multispectral remote sensing data.