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-W5-407-2017
https://doi.org/10.5194/isprs-archives-XLII-2-W5-407-2017
18 Aug 2017
 | 18 Aug 2017

INTEGRATION OF POINT CLOUDS AND IMAGES ACQUIRED FROM A LOW-COST NIR CAMERA SENSOR FOR CULTURAL HERITAGE PURPOSES

M. Kedzierski, P. Walczykowski, M. Wojtkowska, and A. Fryskowska

Keywords: integration, point cloud, NIR images, transformation, texturing, mapping, Old Town in Warsaw

Abstract. Terrestrial Laser Scanning is currently one of the most common techniques for modelling and documenting structures of cultural heritage. However, only geometric information on its own, without the addition of imagery data is insufficient when formulating a precise statement about the status of studies structure, for feature extraction or indicating the sites to be restored. Therefore, the Authors propose the integration of spatial data from terrestrial laser scanning with imaging data from low-cost cameras. The use of images from low-cost cameras makes it possible to limit the costs needed to complete such a study, and thus, increasing the possibility of intensifying the frequency of photographing and monitoring of the given structure. As a result, the analysed cultural heritage structures can be monitored more closely and in more detail, meaning that the technical documentation concerning this structure is also more precise. To supplement the laser scanning information, the Authors propose using both images taken both in the near-infrared range and in the visible range. This choice is motivated by the fact that not all important features of historical structures are always visible RGB, but they can be identified in NIR imagery, which, with the additional merging with a three-dimensional point cloud, gives full spatial information about the cultural heritage structure in question. The Authors proposed an algorithm that automates the process of integrating NIR images with a point cloud using parameters, which had been calculated during the transformation of RGB images. A number of conditions affecting the accuracy of the texturing had been studies, in particular, the impact of the geometry of the distribution of adjustment points and their amount on the accuracy of the integration process, the correlation between the intensity value and the error on specific points using images in different ranges of the electromagnetic spectrum and the selection of the optimal method of transforming the acquired imagery. As a result of the research, an innovative solution was achieved, giving high accuracy results and taking into account a number of factors important in the creation of the documentation of historical structures. In addition, thanks to the designed algorithm, the final result can be obtained in a very short time at a high level of automation, in relation to similar types of studies, meaning that it would be possible to obtain a significant data set for further analyses and more detailed monitoring of the state of the historical structures.