Texture in high-resolution satellite images requires substantial amendment in the conventional segmentation algorithms. A measure is proposed to compute the Hölder exponent (HE) to assess the roughness or smoothness around each pixel of the image. The localized singularity information is incorporated in computing the HE. An optimum window size is evaluated so that HE reacts to localized singularity. A two-step iterative procedure for clustering the transformed HE image is adapted to identify the range of HE, densely occupied in the kernel and to partition Hölder exponents into a cluster that matches with the range. Hölder exponent values (noise or not associated with the other cluster) are clubbed to a nearest possible cluster using the local maximum likelihood analysis.