Detecting the features of urban areas in detail requires very high spatial and spectral resolution in images. Hyperspectral sensors usually offer high spectral resolution images with a low spatial resolution. By contrast, multispectral sensors produce high spatial resolution images with a poor spectral resolution. Therefore, numerous fusion algorithms and techniques have been proposed in recent years to obtain high-quality images with improved spatial and spectral resolutions by sensibly combining the data acquired for the same scene. This work aims to exploit the extracted information from images in an effective way. To achieve this objective, a new algorithm based on transformation was developed. This algorithm primarily depends on the Gram–Schmidt process for fusing images, removing distortions, and improving the appearance of images. Images are first fused by using the Gram–Schmidt pansharpening method. The obtained fused image is utilized in the classification process in different areas by using support vector machine (SVM). The classification result is evaluated using a matrix of errors. The overall accuracy produced from the hyperspectral, multispectral and fused images was 72.33%, 82.83%, and 89.34%, respectively. Results showed that the developed algorithm improved the image enhancement and image fusion. Moreover, the developed algorithm has the ability to produce an imaging product with high spatial resolution and high-quality spectral data.