The Virtual Sculptures and Monuments project has begun as a practical usage of data obtained for my master’s thesis in the field of Land Surveying and Cartography. Titled Comparison of Dense Stereo Matching and Laser Scanning on the Example of Chosen Sculptures and Monuments, a thesis was finished and defended in the year 2014 on Technical University of Koszalin. The master’s thesis was supervised by Wolfgang Kresse, PhD. Eng., who I wish to thank, especially for lending me a terrestrial laser scanner (FARO Laser Scanner Focus3D X 130) from German Hochschule Neubrandenburg.
The thesis wasn’t focused on a spacial reconstruction itself. Instead, as the subject indicates, it was an analysis of advantages and limitations of two measuring methods used for a reconstruction on the example of different objects related to a cultural and national heritage. In the result I have obtained a valuable data, at least in regard of the cost of used equipment and software. The data could promote local region (of city of Koszalin) and original authors of sculptures and monuments. If any of these objects is damaged or destroyed, it’s virtual copy will persist.
Unfortunately, regardless of my inquires, the obtained spacial data haven’t found a special interest. The idea of website creating the possibility to watch 3D models is the practical use of these data. The implementation of this idea started at the end of year 2017. I can also add, that during my visit in the National Archive in Koszalin I mentioned about the data of sculptures and monuments. In the result I received the contact to people which possibly are interested in, thus the data could appear in further publications.
The project is realized as hobby, yet creates financial costs. Thus my personal request to you for support and the funding of this project. Each cent will matter.
Generalities of reconstruction
There are two kinds of 3D models on this website. The one that are created from terrestrial laser scanning and the second that originate from spacial reconstruction of multiple stereo pairs of photos. Both methods have their advantages and disadvantages, complementary to some extent when combined. The result is a highly detailed model weighting 800 Megabytes. This size is still to large to be casually displayed via Internet. Thus, the model is optimized to the size around 10 MB and may be manually corrected. Each model view downloads about 10 MB, which should be remembered while using low quality Internet connection and especially when using pay per use bandwidth. The exact sizes of files are displayed for each object.
Measuring (limitation of used technology, properties of object and scene), combining and alignment, reducing the model and texturing - these are steps of reconstruction and each of it creates, carries and multiplies errors visible in the final model. Because of using various complicated algorithms conducting mass calculations, the final error is difficult to estimate. The detailed metric of measured objects is not in focus of this project due to the fact of not having a direct access to terrestrial laser scanner. Instead, the website is aimed to public and free of charge presentation of reconstructed models in the form working on Internet browsers.
What is next?
The plans are to improve already existing models and to reconstruct other sculptures and monuments placed in Koszalin and around. I’m also considering reconstruction of buildings and locations. However I am limited as the full-time employee. Also, the local weather does not always allow proper photographic session in 3D reconstruction purpose.
Because I’m in possession of commercial license for the 3D reconstruction software , I can also create other similar models of various objects, of which geometry, scene conditions and materials make photo-based reconstruction possible. For such could be considered items and even persons. Received model could be reprocessed to 3D printable form. The original master thesis in Polish language is also purchasable, in example on this website pracemagisterskie.pl.