Validation of Infrared Scanner by the Assistance of Geomatic Documentation of the Historical Building of Etemad al-Saltanah

Since the historical buildings undergo a lot of changes and damages in the course of history, they are required to be documented. These changes might come about as a result of natural symptoms like rainfall, wind, earthquake, flood, explosion and/or by human beings (consciously or unconsciously). Therefore, efforts should be made in line with 3D documenting such buildings so that, besides precise identification of the buildings’ current status and the damages imposed to them in the past, the future changes and damages’ trends could be predicted so as to prevent their continuation. The documentation system is selected according to the dimensions of the object, the density of the required point clouds and accuracy. Regarding that the current methods for laser-based or photography-based (photogrammetry) 3D reconstruction are expensive or complex, cost-effective infrared sensors, such as the structure sensor and the Kinect sensor, have been introduced as promising alternative tools. An infrared scanner, as a portable depth-sensing scanner, consists of a color sensor and a depth sensor that are capable of capturing color images and depths of objects in the visible and accessible range. These sensors are commonly referred to as RGB-D cameras because they output standard RGB images from the camera that have an additional Depth channel per pixel (Fig. 2). The most recent development of the infrared documentation system is the portable Structural Sensor provided by Occipital in collaboration with Prime Sense. This small, lightweight, wireless sensor directly collects and records point clouds data and create three-dimensional modeling of interiors. Structure sensor is a new technology in metric documentation, therefore, the capabilities of this system have not been evaluated for documenting cultural heritage. According to the error introduced for the structure sensor, the scanner has a precision of more than 99% in objects between 0.4 and 3.5 meters; therefore, it is suitable for heritage documentation. The main purpose of this research is, therefore, to verify this claim based on projects captured through experimental tests, in order to confirm the suitability of this tool for cultural heritage documentation.  The historical house of Etemad al-Saltanah was documented (Fig. 8) and processed to experimentally examine the structure sensor, the results of which were compared with the actual dimensions of the house (Table 4). Results of the research show that this system of documentation is not suitable for 3D capturing and reconstructing historical buildings and does not have the required and claimed level of precision (Table 5). Also, the structure sensor precision was assessed for documenting museum objects through testing the scale model of Imam Mosque in Isfahan, Iran (Fig. 11). Results (Table 6) indicate that the structure sensor is only suitable for historical objects with dimensions between 0.3 and 2 meters, and has a precision of more than 95%, which is acceptable according to the Cadastral spatial information regulation. The number of point clouds varies between 103 and 106 points in each capture (Fig. 12) and the capture dimensions are achievable considering a root-mean-square error up to 5 m3, beyond which is higher than the capability of the scanner. Pearson correlation test shows increasing errors of the scanner with enlarged sizes of objects (Table 3).