<p>Fused Filament Fabrication (FFF) enables rapid and flexible manufacturing of complex geometries, but the layer-by-layer deposition process inherently introduces defects such as voids and geometric deviations. This study investigates the influence of raster deposition configuration and deposition order on the mesoscale geometry of 3D-printed PLA samples. Samples with aligned and shifted 3 × 3 raster configurations were fabricated using open-source FullControl G-code generation and analyzed with micro-computed tomography (µCT). Image segmentation methods, including watershed segmentation and curvature-based analysis, were used to extract void geometries and raster boundaries for subsequent computer-aided design (CAD) reconstruction. Results show that void morphology and raster cross-sectional areas are significantly affected by both configuration and deposition order. In aligned samples, voids tended toward diamond-shaped geometries, whereas shifted samples produced triangular voids; however, actual geometries frequently deviated from these idealized forms, leading to unexpected connections and larger voids. Deposition order was found to influence void size by as much as 30–45%, with configuration sequences producing larger voids. Raster analysis further revealed that thermal history and raster sequence altered cross-sectional areas, with final rasters in a row exhibiting larger dimensions. These findings demonstrate that deposition order and raster configuration are critical parameters governing void content and raster uniformity, both of which are directly linked to mechanical performance. The presented methodology provides practical insight for optimizing G-code generation and adaptive process planning to minimize void formation and enhance print quality. The integration of µCT-based analysis and CAD reconstruction establishes a foundation for developing high-fidelity digital twins and predictive models of FFF-printed structures.</p>