ABSTRACT:This study aims to compare classification accuracies of land cover/use maps created from Sentinel-2 and Landsat-8 data. Istanbul metropolitan city of Turkey, with a population of around 14 million, having different landscape characteristics was selected as study area. Water, forest, agricultural areas, grasslands, transport network, urban, airport-industrial units and barren land-mine land cover/use classes adapted from CORINE nomenclature were used as main land cover/use classes to identify. To fulfil the aims of this research, recently acquired dated 08/02/2016 Sentinel-2 and dated 22/02/2016 Landsat-8 images of Istanbul were obtained and image pre-processing steps like atmospheric and geometric correction were employed. Both Sentinel-2 and Landsat-8 images were resampled to 30m pixel size after geometric correction and similar spectral bands for both satellites were selected to create a similar base for these multi-sensor data. Maximum Likelihood (MLC) and Support Vector Machine (SVM) supervised classification methods were applied to both data sets to accurately identify eight different land cover/ use classes. Error matrix was created using same reference points for Sentinel-2 and Landsat-8 classifications. After the classification accuracy, results were compared to find out the best approach to create current land cover/use map of the region. The results of MLC and SVM classification methods were compared for both images.

This research aims to investigate how different landscape metrics are affected by the enhancement of the thematic classes in land cover/land use (LC/LU) maps. For this aim, three different LC/LU maps based on three different levels of CORINE (Coordination of Information on The Environment) nomenclature were created for the selected study area using GEOBIA (Geographic Object Based Image Analysis) techniques. First, second and third level LC/LU maps of the study area have five, thirteen and twenty-seven hierarchical thematic classes, respectively. High-resolution Spot 7 images with 1.5 m spatial resolution were used as the main Earth Observation data to create LC/LU maps. Additional geospatial data from open sources (OpenStreetMap and Wikimapia) were also integrated to the classification in order to identify some of the 2nd and 3rd level LC/LU classes. Classification procedure was initially conducted for Level 3 classes in which we developed decision trees to be used in object-based classification. Afterwards, Level 3 classes were merged to create Level 2 LC/LU map and then Level 2 classes were merged to create the Level 1 LC/LU map according to CORINE nomenclature. The accuracy of Level 1, Level 2, Level 3 maps are calculated as; 93.50%, 89.00%, 85.50% respectively. At the last stage, several landscape metrics such as Number of Patch (NP), Edge Density (ED), Largest Patch Index (LPI), Euclidean Nearest Neighbor Distance (ENN), Splitting Index (SPLIT) and Aggregation Index (AI) metrics and others were calculated for different level LC/LU maps and landscape metrics values were compared to analyze the impact of changing thematic details on landscape metrics. Our results show that, increasing the thematic detail allows landscape characteristics to be defined more precisely and ensure comprehensive assessment of cause and effect relationships between classes.

ABSTRACT:This study aims to compare classification accuracies of land cover/use maps created from Sentinel-2 and Landsat-8 data. Istanbul metropolitan city of Turkey, with a population of around 14 million, having different landscape characteristics was selected as study area. Water, forest, agricultural areas, grasslands, transport network, urban, airport-industrial units and barren land-mine land cover/use classes adapted from CORINE nomenclature were used as main land cover/use classes to identify. To fulfil the aims of this research, recently acquired dated 08/02/2016 Sentinel-2 and dated 22/02/2016 Landsat-8 images of Istanbul were obtained and image pre-processing steps like atmospheric and geometric correction were employed. Both Sentinel-2 and Landsat-8 images were resampled to 30m pixel size after geometric correction and similar spectral bands for both satellites were selected to create a similar base for these multi-sensor data. Maximum Likelihood (MLC) and Support Vector Machine (SVM) supervised classification methods were applied to both data sets to accurately identify eight different land cover/ use classes. Error matrix was created using same reference points for Sentinel-2 and Landsat-8 classifications. After the classification accuracy, results were compared to find out the best approach to create current land cover/use map of the region. The results of MLC and SVM classification methods were compared for both images.

No abstract