Heterogeneity is an inherent characteristic of most engineering materials, manifesting across multiple length scales. At the millimetre scale, for example, the composite nature of concrete is visually apparent, whereas the grain-level heterogeneity of metallic alloys typically requires high‑resolution techniques such as Electron Backscatter Diffraction (EBSD) for observation at the micrometre scale. Materials such as rocks, composites, and alloys comprise multiple phases distributed randomly throughout their domains, and their macroscopic behaviour is strongly governed by the morphology and spatial arrangement of these microstructural features. Characterisation of such microstructures commonly relies on imaging techniques including Scanning Electron Microscopy (SEM), micro‑Computed Tomography (micro‑CT), and optical microscopy. Due to their intrinsic randomness, heterogeneous microstructures are often described using statistical descriptors that capture key morphological and geometrical attributes. While these descriptors are essential for classification and comparison, they provide limited capability for quantitative prediction of material properties. Accurate property prediction requires digital reconstruction of the microstructure, for which a range of reconstruction techniques has been developed across different material classes. Once a representative digital microstructure is obtained, physics‑based simulations can be performed to evaluate effective macroscopic properties through multiscale modelling approaches, including Representative Volume Element (RVE) analysis. This presentation provides a comprehensive overview of the theoretical foundations, computational algorithms, and practical applications associated with microstructural characterisation, digital reconstruction, and Uncertainty Quantification (UQ) of heterogeneous materials. Together, these components form an integrated framework for understanding and predicting the behaviour of complex material systems.