The first two years of the project were dedicated to the collection, assessment and analysis of historic sources and the development of a novel methodology for on-site 3D documentation of window frames from 1960 to 1990. The project is based on the assumption that the information available in historic sources is sufficient for identification of samples with sufficient accuracy and with high efficiency, preferably semi-automatically. The objective is to compare the profiles and types of window frames found in the on-site investigations (Fig. 1) with data from historic sources. In accordance with the project schedule, the initial focus was on objects whose aluminium framed windows are in good condition and whose manufacturers are known.
Besides terrestrial laser scanning and 3D photogrammetry, a recording method for precision measurements has been developed using a handheld blue light laser scanner which is attached to a measuring arm. The approach was first tested and refined using approximately two dozen mock-ups of window frames under stable laboratory conditions in-house. It was then applied on-site in a total of eight buildings so far, where a total of 108 window frames were scanned, resulting in 216 individual scans and 3D models.
Even though the study focuses on manufacturers of windows, window profiles, and facades in German-speaking countries, the historical analysis is confronted with an almost overwhelming abundance of products, types, series, brand names, property rights, and patents that were marketed in contemporary publications (Fig. 2). An essential basis for the project are product catalogues available in analogue form (Fig. 3). These are currently evaluated, compared, and digitized as part of the project. Alphanumeric and geometric in-formation on the profiles is brought together in IFC4 data models. To date, the project’s collection of identified and assessed products comprises over 380 window systems and types from more than 55 manufacturers, with the current focus, in line with the project schedule, on aluminium frames. The corresponding study of plastic (PVC), steel, and composite frames has been started. In particular, the support of manufacturers and respective industrial associations has been secured for the project, as well as access to several company archives.
In order to achieve the project objective, it will be critical to reconcile the data collected from archival and literature research with the properties of the objects as determined on site. The measurement data must not be rendered unidentifiable due to poor condition or inaccuracies in the recordings. The aim of the project is to identify existing window systems with sufficient accuracy on the basis of targeted and limited visual inspection using the database. We have developed a novel methodology to validate our 3D scans against ground truth (GT) data that was generated using paper-based historic plans. These plans are digitized and then converted into 3D data. They are compared to the scans that we acquired on-site (Fig. 4). The data acquisition and 3D simulation method have been refined to such an extent that measurement inaccuracies or imprecisions in mesh processing can be ruled out with a high degree of probability in the case of detectable deviations between GT and scan. The inaccuracies that can be measured and verified in high resolution now allow conclusions to be drawn about, for example, manufacturing processes, but also maintenance and usage intensities. A peer-reviewed publication for this important step of our research has been approved for publication.
As part of the accurate capture and realistic visualization of the objects in question, we have worked on replicating their real-world appearance. This step involved the use 3D laser scanning and photogrammetry, and will involve their fusion as well as the integration of material and reflection properties of individual objects (Fig. 5). We also aim to bring these objects into a 3D computer graphics platform to allow for their high-fidelity (virtual) analysis for high resolution condition studies.
