earticle

영어

This study focuses on enhancing the diagnostic capabilities of junior fire investigators through the optimization of AI microscope measurement methods for identifying wire arc melting marks. A total of 33 wire samples collected from actual fire scenes were analyzed using three distinct imaging scenarios (S1, S2, and S3) to evaluate the most reliable method. The S1 scenario, which involved capturing the wire’s side with the widest distribution of melting marks, achieved the highest concordance rate of 90.9% with metallographic analysis. In contrast, S2 and S3 scenarios recorded lower concordance rates of 51.5% and 39.4%, respectively. Statistical analysis using ANOVA revealed significant differences among the methods (F = 9.80, p < 0.00014), with post-hoc tests confirming the superior accuracy of the S1 approach. These findings suggest that AI microscopes can effectively support junior fire investigators by improving diagnostic accuracy and reducing reliance on subjective judgment.