The XXI Congress of the International Society for Photogrammetry and Remote Sensing (ISPRS 2008), July 3-11, 2008, Beijing, China
cave; cultural heritage; descriptor; fresco; indexing; invariant; modelling; reconstruction; restoration
This paper describes a computer-aided method for the reconstruction and modelling of complex sites based on content-based indexing of images. A pose, distortion and composition robust descriptor is automatically generated for each image associated with a given site. These descriptors are subsequently placed in a database, to be used for future querying. The database is interrogated in real-time using an intuitive query by example paradigm, providing visual results which may be easily understood and interpreted. Our approach is applied to the modelling of the Grotta dei Cervi and to the restoration of the broken frescoes of the Assisi Basilica.
The XXI Congress of the International Society for Photogrammetry and Remote Sensing (ISPRS 2008) [Proceedings].