In this research, water body characterization is performed by decomposing the water bodies into convex components using morphological decomposition. Opening by reconstruction is implemented on the decomposed convex components using square kernels of increasing sizes. A power law relationship is observed between the number of decomposed convex components removed at each iteration of opening by reconstruction and the kernel size. The scaling exponent of this power law is the fractal dimension of the water bodies, which indicates the measure of complexity of the self-similarity of the water bodies. The iterative opening by reconstruction process is implemented on the individual water bodies to compute their fractal dimensions. The computed fractal dimensions are shape dependant and consider the topological regions of water bodies rather than their geometric boundaries.