Image segmentation is the process of dividing the image into semantically significant regions or objects. These regions are recognized by further processing steps. This project represents a novel spatially-constrained color-texture model for hierarchical segmentation of very high resolution images. In this project we are segmenting the given image into meaningful regions where we can extract the required regions for future uses such as content based image retrieval, video surveillance, medical imaging and in so many areas. For a given image, the method starts with initial partition where the image is partitioned into many homogenous regions, there after represent the regions using a region adjacency graph in which a novel spatially-constrained color-texture model is used to measure the distances between adjacent regions. Finally a step wise optimized region merging process is applied to obtain hierarchical segmentation results. These results show that the proposed method is highly efficient and gives better results among other existing color–texture segmentation methods. The areas of different colors and textures are properly partitioned. That is partially contributed by the adaptive weighting of color and textural features, which leads to the full use of the color and texture features. The spatial constraint contributes a lot to the good performance.