This study investigates the influence of particle shape on the void space morphology and topology in granular materials. Numerical samples with spherical and ellipsoidal particle shapes were generated using the discrete element method. A segmentation algorithm was used to extract the pore space characteristics. The results reveal that particle shape significantly affects both constriction and pore sizes, with distinctive features according to flatness index or elongation ratio, the former being more significant than the latter. The obtained results were validated by conducting numerical filtration tests, which illustrated a direct correlation between the constriction properties derived from the pore space extraction and the blockage rate of fine particles in the filtration tests. The study revealed the importance of considering particle shape in filter design, emphasising its significant impact on pore space characteristics and filtration performance.