![<small>(Image: Grant Recipient or NCPTT)</small> 2004-01](https://webarchive.library.unt.edu/eot2008/20090114203231im_/http://www.ncptt.nps.gov/images/Product-Catalog/250/2004-01-250.jpg)
Traditional thin-section analysis techniques such as tedious manual point counting are time consuming and require numerous measures. The use of visual estimation charts produce results of questionable accuracy. The recent advent of general purpose digital image analysis software, ubiquitous computing, and digital microscope cameras make rapid analysis commonly accessible. This study examines the utility of a trio of software packages for analysis of petrographic thin sections in conservation research. In addition to spatial calibration considerations the study recommends a set of protocols measuring microcracks in quartz grains, phase measurement, and layer-thickness measurement. The protocols are designed to help researchers choose a package and begin work at once.
Although none of the packages under review could identify crack patterns, all were able to readily measure the length of microcracks in quartz grains and two were able to provide minimum, maximum, and average widths. However, each required user intervention to remove features misidentified as cracks and unwanted trace lines; under most circumstances manually tracing cracks with a mouse provided the best result.
Layer thickness measurement provides crucial indicators used to determine the degree of weathering in cultural artifacts and the techniques and technologies used in artifact fabrication. The chief limitation of layer thickness analysis remains its reliance on the human eye to determine layer boundaries. Software which ameliorates this problem by using gray and color threshold analysis to separate layers requires highly variable and complex measurement and analysis procedures, hampering its utility in analyzing numerous samples.
Determining the area percent, size, and shape of phases in ceramic thin sections requires numerous measurements taken from random locations in the sample and, ideally, multiple section samples both perpendicular and parallel. Though each tool examined provides features to automatically analyze these traits, some samples may need to have matter manually identified where the software cannot correctly identify boundaries between the matrix and particles under study based on an analysis of gray-scale or color gradations.
Although each package varies in ease of use and utility for a given application, these tools supplant the time consuming traditional methods of petrographic thin section analysis that dominated conservation research. Researchers can obtain more accurate and timely results through the application of these technologies according to the protocols outlined in this study.