Use of Sentinel-2 for land cover classification on Shar Planina, North Macedonia

Providing information on the spatial distribution of habitat groups through land cover classification is useful for a wide range of environmental studies, especially those monitoring and managing succession and land cover change. While Corine Land Cover is widely used in standardised land cover comparisons, its strength is primarily its use in coarse scale analyses over large extent. Recently available Sentinel-2 combined with wide availability of open-source remote sensing software allows customised fine-scale land cover classifications to suit research projects applied over a smaller extent. In this regard, this study aims to test the suitability of Sentinel-2 imagery in providing a customised land cover classification for Shar Planina mountain range in North Macedonia, hoping that the results will serve as a background for assessment of its landscape visual quality as part of the ongoing PhD study (Jovanovska, D. 2018-2021). The results show that use of Sentinel-2 multiband high resolution images allows customized medium-large scale image classification that yields great accuracy results. Furthermore, frequent updating of Sentinel-2 database allows comparability and standardised land cover change analysis for the purpose of environmental monitoring. The resulting land cover data classification is uploaded as an .ASCII file accompanied with corresponding .lyr file and can serve as a background for subsequent environmental studies in the region. Finally, any further improvements of the final output, as well as trials on its suitability in studies with different aim and scope are encouraged and welcomed.通过土地覆盖分类（land cover classification）获取生境群的空间分布信息，可支撑各类环境研究，尤其是聚焦生态演替与土地覆盖变化的监测与管理类研究。科里宁土地覆盖（Corine Land Cover）数据集虽广泛应用于标准化土地覆盖对比研究，但其核心优势主要体现在大尺度范围的粗分辨率分析场景中。近年来，哨兵二号（Sentinel-2）卫星数据的普及，配合开源遥感软件的广泛可用，使得针对小尺度研究项目定制化的精细尺度土地覆盖分类方案成为可能。有鉴于此，本研究旨在验证哨兵二号（Sentinel-2）影像应用于北马其顿沙尔山脉（Shar Planina）定制化土地覆盖分类的适用性，以期研究结果可为该区域景观视觉质量评估提供基础背景数据，本研究作为正在进行的博士研究的一部分（约万诺夫斯卡，D. 2018-2021）。研究结果显示，利用哨兵二号多波段高分辨率影像可实现定制化的中大规模影像分类，且分类精度优异。此外，哨兵二号数据库的高频更新特性，可为环境监测场景下的土地覆盖变化对比分析与标准化研究提供可靠支撑。本研究生成的土地覆盖数据分类结果以.ASCII文件格式上传，并附带对应的.lyr文件，可作为该区域后续环境研究的基础背景数据。最后，本研究鼓励并欢迎对最终输出结果开展进一步优化，以及针对其在不同目标与研究范围中的适用性开展相关测试。