Surface analysis of Roman Sarcophagus using multispectral close-range photogrammetry

Currently, cameras equipped with multispectral sensors are utilized to monitor cultural heritage. However, these cameras have a lower resolution compared to RGB cameras, resulting in a loss of detail during analysis. The method proposed in this study outlines a series of image processing and data analysis techniques for orthophotos generated through the photogrammetric evaluation of images captured with RGB and multispectral cameras. This study first enhances the low-resolution multispectral orthophotos and creates a multispectral image stack. The resulting image stack undergoes dimensional reduction via principal component analysis, followed by the implementation of a clustering algorithm. Consequently, a map is generated to identify and examine surface details using multispectral data. The discrimination capability was verified comparatively by analyzing the histograms of selected subregions from the generated maps. This validation demonstrates that meaningful information is successfully produced from the image stacks. This method offers a rapid, efficient, and adaptable analysis of photogrammetric products necessary for the field of cultural heritage.