Motivation

The surfaces of all cultural assets located outdoors are more or less exposed to atmospheric influences, which can lead to damage if no protective measures are taken or these do not have the desired effect. With regard to rock surfaces, mainly physical damage can be observed, which leads to the erosion of the rock substance. Early detection of the type, distribution and extent of such damage is an important prerequisite for taking targeted and adequate countermeasures that are intended to preserve the valuable object substance. In view of the gradual characteristics of such decay processes, a qualitative and quantitative assessment is problematic, since only minimal changes often occur in short time intervals. On the other hand, there is a good chance of preserving the surfaces, especially for slow processes, if the processes running can be recognized and evaluated at an early stage. However, this requires a very detailed analysis of the surface and the morphological changes taking place therein. The basis for such analyzes must be very precise geometric and, if necessary, color or spectrometric measurements that document the morphological and optical condition of the surfaces and provide a reliable data basis for the detection of changes.

Activities

In the context of area monitoring, the Institute for Stone Conservation and the Institute for Spatial Information and Measurement Technology (i3mainz) work together to explore the potential of modern measurement methods and use them to monitor selected object areas. In a first phase, the potential of various measurement methods was analyzed and characterized in order to be able to select the most suitable technology for the exemplary application that follows in the second phase. The test area at the monastery ruins of Limburg near Bad Dürkheim, which was set up in the 1990s and prepared using geometric references, served as a test object for the process analysis (see Fig. 1).

The second project phase serves to apply the measurement technology defined in the first phase to other objects that are subject to more or less severe decay processes in different ways and caused by various factors. The selected objects (see Table 1) each have different characteristics and were recorded according to the previously defined requirements.

In addition to the metrological requirements regarding the required object resolution, the individual measurement campaigns on an object must be referenced in the same coordinate system. The demands placed on the geometric reference increase with the precision of the measurement process and can currently only be met by appropriate permanently stable markings on the object. However, the necessary interventions on the object represent a critical measure that is rather undesirable from a conservation point of view. For this reason, a decision must be made on a task- and property-related basis as to which marketing measures are possible without unacceptably damaging the property. However, if the established geometric reference has to be at a greater distance from the object to be monitored, this means more effort for the individual measurement campaign and may also entail an additional risk for the achievable geometric quality.

Results

The basis for the process comparison are the results obtained in two measurement campaigns on the test area of ​​the Limburg monastery ruins (see Fig. 1), which are differentiated according to purely geometric, procedural and more practical aspects. Table 2 provides an overview of the geometric aspects. It can be seen that comparable resolutions and accuracies can be achieved thanks to the flexibility of the measurement concepts. However, the way it comes about and the result differ considerably in some cases. The active systems (scanning, projection) provide an explicit description of the surface through a very large number of 3D points, while the passively working photogrammetry only provides an implicit, virtual model. Even after the images have been oriented, the geometric information is still in the images and must first be extracted by means of a further interactive measurement. For a full-surface modeling, as in the case of the active systems, this would involve considerable effort, since a measurement would have to be made for each 3D point. It is also possible to use calculation algorithms, although these have weaknesses in regions with little texture and in strong geometric changes such as edges. The lighting conditions prevailing on the object also have a major influence on the result. But the two active systems also provide different data bases. With high-resolution scanning, the 3D model is assembled from a much larger number of individual measurement steps, which means that on the one hand there is greater local effort, but on the other hand the risk of error when assembling many individual measurements fundamentally increases. The advantage of hand-guided scanning systems is certainly the greater flexibility with more spatially structured surfaces. In this way, the scan head attached to the measuring arm can also be guided into areas with undercuts or into hidden zones. For a projection system, on the other hand, more complex surfaces lead to significantly greater effort, since further measurement arrangements may also be required for only very small areas.

Differences between fringe projection and scanning can be found in Table 3. Additional criteria are listed here that can be used to evaluate the procedures. Here, too, hand-guided scanning performs somewhat poorly because, in addition to the geometric validity of the data, the process is also more complex and difficult to handle. With regard to the explicit creation of two-dimensional 3D models, stripe projection offers the most powerful concept of all methods. The assessment can be different if the surface is not to be recorded over a large area, but rather individually and selectively. In such cases, a passive method such as stereophotogrammetry can demonstrate its advantages. It is easy to use, the equipment is inexpensive and images also provide direct visual information about the surface. Photogrammetry can also be easier to handle with regard to possible interventions and preparations on the object surface, because the geometric connection between different individual measurements can be made via natural surface features and thus requires less intervention on the object. In this respect, stereophotogrammetry also offers a useful concept, but is to be understood more as a powerful data store from which measurements can be taken selectively if required. With regard to thorough monitoring, which must also detect and explicitly quantify smaller surface changes, the comparison of the alternative methods only leads to the conclusion that fringe projection should be used as a basis for further investigations.

In the second phase of the project, the above concept of fringe projection was applied to four different object surfaces in order to be able to observe and analyze the changes occurring there from different perspectives