<p>Calcium carbonate distribution during the enzyme-induced carbonate precipitation (EICP) process is crucial for strength development, showing both microscopic manifestation of the treatment effect and macroscopic strength development over time. In light of this, a data-driven image characterization method termed “polarization measurement” was employed in this research to study the distribution of calcium carbonate in EICP-treated silica sand by leveraging the differing polarization properties of the materials. Polarization optical techniques combined with an unsupervised k-means algorithm were employed to identify calcium carbonate distribution in specimens subjected to different numbers of treatment cycles. Additionally, miniature cone penetration tests quantified the effects of calcium carbonate content and accumulation modes on sample strength. Results indicate that the k-means method, based on 16.72 million pixels of polarization data, can effectively characterize the distribution of CaCO₃. Compared with the SEM results, the Fourier descriptor cosine similarity exceeds 0.90,&#xa0;with a corresponding average Euclidean distance of less than 0.09, and the results are consistent with EDX analysis. Polarization analysis shows that, as the number of treatment cycles increases, the CaCO₃ content increased from 10.18% to approximately 37.82%, with peak cone penetration resistance rising by nearly a factor of four, indicating significant strength improvement. SEM observations show that the accumulation mode transitioned from particle bridging to surface coating and eventually to overall interconnection, greatly enhancing strength. This is the novel application of polarization measurement in EICP, highlighting its potential in geotechnical characterization.</p>