This study presents a methodology for segmenting, analyzing and simulating at pore scale high-resolution CT scanned images of heterogeneous samples, focusing on an SCR Diesel Particulate Filter (SDPF). Achieving a voxel resolution of approximately 360 nm enabled the identification of small voids within the washcoat, non-resolved with lower resolution. These voids could affect main filtration characteristics and should be accounted carefully for the transport simulations at the pore scale. Due to the computational expense of directly simulating over 2.3 billion voxels, in the current work, representative subdomains were selected by averaging grayscale values along the flow direction to reflect the full domain’s properties. However, significant computational resources were still required for these elements. To mitigate this, a downsampling method defined by the downsampling factor (DSF) was introduced, effectively reducing domain size while preserving key features of the sample. Results showed that downsampling substantially lowers computational demands, allowing faster simulations with satisfactory accuracy. For instance, simulations on downsampled domains with DSF = 2 and 4 closely matched those on the original domain, proving the method’s effectiveness. This study underscores the necessity for tailored simulation strategies for high-resolution and heterogeneous samples. The downsampling method not only accelerates simulation processes but also maintains essential sample characteristics, offering a practical solution for full-domain estimations.