U.S. flag

An official website of the United States government, Department of Justice.

Improving Investigative Lead Information and Evidential Significance Assessment for Automotive Paint and the PDQ Database

NCJ Number
Date Published
February 2014
102 pages
The main goal of this project was to develop search pre-filters and library searching algorithms for clear-coat paint spectra in the Paint Data Query (PCQ) automotive paint database.
This effort is needed because modern automotive paints use thinner undercoat and color coat layers protected by a thicker clear-coat layer. As a result, a clear-coat paint smear is sometimes the only layer of automotive paint left at a crime scene. In such cases, the text-based portion of the PDQ database cannot identify the motor vehicle because of the search's reliance on large variations in color and chemical formulations that do not exist with clear coats. Still, clear-coat paint layers, like the undercoat and color coat paint layers, exhibit chemical features in their infrared (IR) spectra that are indicative of the automobile manufacturing plant at which they were applied; therefore clear-coat spectra may be used to identify the model and line of motor vehicle. Currently, the capability to perform direct searching of IR spectra in PDQ does not exist, and spectral search algorithms commercially available cannot distinguish the subtle differences between clear-coat paint spectra from one vehicle model to the next. In order to address this problem, the current project developed a prototype library search system that identifies the manufacturing plant, model, and line of an automobile from its clear-coat paint spectrum. The system consists of two separate but interrelated components: search pre-filters to cull the library spectra to a specific plant or plants and a cross-correlation search algorithm that identifies spectra most similar to the unknown in the set identified by the search pre-filters. As the size of the library is culled for a specific match, the use of the search pre-filters increases both the selectivity and accuracy of the search. 63 figures, 28 tables, and 52 references

Date Published: February 1, 2014