Integration of Sentinel-1A and Sentinel-2B Data for Land Use and Land Cover Mapping of the Kirkuk Governorate, Iraq

Abstract

Land use and land cover maps are essential to aid our knowledge of modelling the environment, managing water. Multispectral and SAR Satellite data consider the main and valuable resource for LULC mapping. Because of the presence of clouds, creating a precise LULC map using multispectral data is a challenge. Herein, the goal of this study is to generate a precise map of LULC of Kirkuk city using different classification methods and to evaluate the impact of combining SAR and optical Sentinel (1A and 2B) data on classification efficiencies. Gram–Schmidt (GS) method was applied to combine the multispectral Sentinel 2B data and Sentinel-1A (VH, VV). The efficiency of using four commonly-used classification algorithms was then compared to specify the optimal method for LULC classification. The finding reveals that the greatest accuracy of 97.93% with a kappa coefficient of 0.97 was produced using the SVM algorithm applied to multispectral Sentinel-2B data. while the DT-KNN algorithm was most efficient when it applied to Sentinel-2B-VH data with an accuracy of 97.60 %. The overall accuracy of RF is also improved when it applied to Sentinel-1A-VV than Sentinel-1A-VH and multispectral data. Additionally, the method developed will be helpful to researchers who continue to use diverse data sources to map various regions. These mapping results represent an essential step toward future soil mapping and mineral estimation.