A comparison of hyperspectral data and worldview-2 images to detect impervious surface

Detection and mapping the impervious surface accurately is one of the important tasks in urban remote sensing. In this study, airborne hyperspectral data and Worldview-2 image were used to classify urban area .The main goal of this study are to compare the hyperspectral data and worldview 2 images and shows the potential of worldview 2 images for detection the impervious surface from the same area. Support vector machine was used as the classification method in both images. The result shows that the hyperspectral data is more accurate for detection of the materials in urban area especially roof type. The overall accuracy is 78% with 0.72 Kappa coefficients but on the other hand the overall accuracy of worldview 2 image is 72% with 0.65 Kappa coefficients. Thus finally based on the result the airborne hyperspectral data is more suitable for detecting the impervious surface in more detail but still there are some limitations. Furthermore the worldview 2 image shows good potential for detection the impervious surface in detail.