LAUSR

ABSTRACT Color information is one of the most important surface features of earthen sites. Different levels of conservative states and diseases affect the results of traditional in-situ color detection on earthen sites. An alternative in-situ testing method is put forward in this article. The reliability and precision of this method is researched in three categories: surface state of preservation, number and distribution of test points, and data analytical methods. The results show that first cracks and holes on the surface of earthen sites reduce data’s accuracy sharply, so these areas on the surface should be eliminated when choosing test points. Second, using more than one test point leads to better typical data results, and minimum density of sampling number is about 16 per 900 cm2. Finally, after outliers are excluded, the regularities of color data’s distribution are analyzed to predict CIELab values’ ranges of such areas, in which way this earthen site’s surface color information is represented. This color detection method is used in two cases presented in the second part of this article.