Research of the state of forests in the precarpathian region by satellite images using the method of supervised classification

Abstract

The purpose of the research is to monitor the forest objects of the Rava-Ruska forestry and determine changes in their areas from 2002 to 2022. The methodology of monitoring the state of the forests of the Rava-Ruska forestry is based on the using of multi-temporal medium-resolution satellite images with their subsequent processing using geographic information systems. For processing, we use images from Landsat 7 (July 2002) and Sentinel-2 (August 2022). We use csupervised classification using the maximum likelihood method. To post-process the classification results, we use Majority Filter. Based on the results, we calculate the area of classes for the relevant years. We perform a comparative analysis of these areas to determine changes in forest objects. Results. The research was conducted on a part of the territory of the Rava-Ruska forestry. Training samples were created for the following objects: deciduous and coniferous forest, fresh deforestation and overgrown deforedtation, open ground and roads, water and agricultural land. Histograms and scatter plots are used to evaluate the training samples. The suoervised classification was carried out using the maximum likelihood method on a part of the forestry territory with further post-processing. As a result of using GIS tools, a comparison of forest changes over two decades was made. The monitoring of the forests of the Rava-Ruska forestry allowed us to identify changes in forest objects, in particular deforestation, which in 2022 amounted to 679.2 hectares. Scientific novelty and practical significance. A methodology for monitoring forests using remote sensing materials has been developed. The possibilities of detecting changes in forest objects using the method of supervised classification are investigated.