Pixel- vs. Object-based Land Use/Land Cover Classification: The Case Study of Mendefera Sub-zone, Eritrea

Abstract

The study compared the performance of pixel- and object-based land use/land cover (LULC) classification for Mendefera sub-zone, Eritrea, using Landsat 8 OLI. Supervised pixel-based image classification was done in ArcMap with Support Vector Machine (SVM) and segmentation object-based image classification in ArcGIS Pro. Post-classification smoothing and high spatial resolution aerial photos along with google earth image were employed to improve the accuracy. DEM and high spectral resolution satellite images were also used in combination with false composite colours during the creation of training samples. Overall accuracies of 83.7% and 67% and Kappa coefficients of 77% and 49% were obtained for pixel- and object-based classifications, respectively. Thus, the study concludes that pixel based LULC classification is the best classification mechanism for the given study area.