Mapping urban land use and land cover variations based on support vector machine algorithm: A case study of Sowme'eh Sara County

Abstract

Land-use change processes present a variety of trajectories depending on local conditions, the regional context, and external forces. Uncontrolled and imprudent alteration of the Land-Cover and Land-Use Change (LCLUC) brings about serious problems in terms of environmental damages. Taking the scale of the LCLUC changes into account, remote sensing technology can be used efficiently and affordability to monitor and detect such large-scale changes. This study is an in-depth analysis of spatial and temporal LCLUC variations in an urban area (Sowme'eh Sara) in Guilan province, northwest Iran. Support Vector Machine (SVM) algorithm and Landsat imageries were utilized to derive the LCLUC variations between 19889 to 2021. The accuracy of the classification was investigated through ground truth data, Google Earth images, and aerial photographs. The results indicate that the SVM-based classification of TM and OLI imageries gives an overall precision of 93%, and 95% respectively which manifests the high feasibility of this technique for deriving LCLUC. According to the mapping results, a distinctive evolution in LCLUC is seen over the study area where consistent urban consolidation with changes of the wetland typology involving marsh degradation, gains from agro-forest land, or sparsely vegetated areas.