The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Download
Publications Copernicus
Download
Citation
Articles | Volume XLII-2/W5
https://doi.org/10.5194/isprs-archives-XLII-2-W5-505-2017
https://doi.org/10.5194/isprs-archives-XLII-2-W5-505-2017
21 Aug 2017
 | 21 Aug 2017

TESTING DIFFERENT SURVEY TECHNIQUES TO MODEL ARCHITECTONIC NARROW SPACES

A. Mandelli, F. Fassi, L. Perfetti, and C. Polari

Keywords: Laser scanner, Photogrammetry, Fish-eye, Panoramic images, SLAM (Simultaneous Localisation And Mapping), Cultural Heritage survey, Narrow spaces, Accuracy

Abstract. In the architectural survey field, there has been the spread of a vast number of automated techniques. However, it is important to underline the gap that exists between the technical specification sheet of a particular instrument and its usability, accuracy and level of automation reachable in real cases scenario, especially speaking about Cultural Heritage (CH) field.

In fact, even if the technical specifications (range, accuracy and field of view) are known for each instrument, their functioning and features are influenced by the environment, shape and materials of the object. The results depend more on how techniques are employed than the nominal specifications of the instruments. The aim of this article is to evaluate the real usability, for the 1:50 architectonic restitution scale, of common and not so common survey techniques applied to the complex scenario of dark, intricate and narrow spaces such as service areas, corridors and stairs of Milan’s cathedral indoors. Tests have shown that the quality of the results is strongly affected by side-issues like the impossibility of following the theoretical ideal methodology when survey such spaces. The tested instruments are: the laser scanner Leica C10, the GeoSLAM ZEB1, the DOT DPI 8 and two photogrammetric setups, a full frame camera with a fisheye lens and the NCTech iSTAR, a panoramic camera. Each instrument presents advantages and limits concerning both the sensors themselves and the acquisition phase.