期刊名称：ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
印刷版ISSN：2194-9042
电子版ISSN：2194-9050
出版年度：2001
卷号：XXXIV-5/C7
出版社：Copernicus Publications
摘要：This paper describes an automatic appr oach in archaeological photogr ammetry for the documentation and recover y of rupestr ian paintings. Visible and near-infrared images were taken from an opened-cave before and after moistening the stone wall. Digital image processing techniques, specially multispectral and multiband classifications, were used to recognise and extract non- destructively the primitive paintings and shapes, i.e. arrows, hunters and animals. Images of rupestrian paintings are presented, before and after human restoration. Additionally, thematic plots of work art obtained semi-automatically and independently before the cave restoration are shown in order to check the profitability of the methodology; statistical accuracies are also presented. Furthermore, some guidelines for future documentation and recovering of rupestrian paintings are pointed out
关键词：archaeological photogrammetr y; spectral classification; multiband classification; r ock art
