Abstract:

The research evaluates the image fusion techniques using RASAT image which is one of the optical remote sensing satellites launched by the Republic of Turkey. The data has 7.5 m ground resolution panchromatic and 15 m multispectral bands (red, green and blue). The aim of the study is to compare the images fusion methods to achieve spatially improved and spectrally preserved higher resolution RASAT data. The performance of the data was investigated by different image fusion methods. For this purpose, Ehlers fusion, High Pass Filtering (HPF), Intensity Hue Saturation (IHS) and Principal Component Analysis (PCA) data fusion methods were applied and investigated. The study area is located in Menemen Izmir province, west of Turkey. The area has different land use classes such as cultivated fields, bareland, wetland, water body and pasture. Qualitative and quantitative analyses were applied to assess the performance of the fused images. Lower resolution multispectral data was compared to fused images visually and color distortions were investigated. For the quantitative analysis different statistical metrics were utilized. In this frame, Correlation Coefficient (CC), Universal Image Quality Index (UIQI), Root Mean Square Deviation (RMSE) and Relative Dimensionless Global Error in Synthesis-Erreur Relative Globale Adimensionnelle de Synthese (ERGAS) were performed for quality assessments of spatially improved data. Regarding to the results of the pan-sharpening methods it is concluded that Ehlers preserved the best color information while the result of HPF provided higher statistical results. 

Keywords: