Computerized Database for Bullet Comparison by Consecutive Matching

The analysis compared the locations of prominent features of the land impression topography. The database that was used to test the method contained 1,686 land impressions fired from Beretta M9/92 pistols and then scanned and processed with confocal microscopy. The confocal microscopes used could distinguish the small variations in depth that are associated with the land impressions. This was supplemented with additional random-simulated data that brought the database to 90,000 bullets. A total of 232 of the 281 scanned bullets were duplicate firings, i.e., pairs of bullets were fired from 116 guns without excluding any bullets because of the quality of the impression marks. Approximately 60 percent of the duplicate firings were identified before a false positive match occurred. The implications of this study pertain to several issues facing the field of firearms examination. It provides the methodology for constructing a workable bullet database. It also provides a method for quantifying the significance of a bullet match. Further research could assist in developing computer software for scanning curved surfaces, which could significantly reduce the data-acquisition times as well as routine for comparing the reconstructed three-dimensional topography. This project is currently capable of scanning approximately 10 bullets a day. At this rate, a database of 10,000 bullets would take approximately a year, using 3 confocal instruments. 24 figures and 22 references