O uso de aeronaves remotamente pilotadas em protocolos de monitoramento, diagnóstico e documentação científica para conservação do patrimônio cultural arquitetônico

Abstract

This work aims primarily to investigate the use of Remotely Piloted Aircraft (RPA) as a tool for scientific documentation, within protocols for monitoring and diagnosing the conservation of architectural cultural heritage. The study begins with a comparative survey of the state of the art in the integrated use of Historic Building Information Modeling (HBIM) technologies worldwide and, more specifically, in Brazil. The proposed case study focuses on monitoring constructive pathological manifestations in the Church of Saint Francis of Assisi, better known as the Igrejinha da Pampulha, an iconic work of Brazilian modern architecture located in Belo Horizonte, MG. The proposed method involves pre-programmed georeferenced flight plans with ground control points, digital image processing, point cloud generation, and the treatment of three-dimensional computational models aimed at identifying and characterizing changes in the conservation state and in the pathologies investigated in different parts of the building. The results encompass the development and testing of monitoring, diagnostic, and scientific documentation protocols that may serve as a basis for decision-making in preventive, curative, and restoration conservation routines and processes in the case studied. The method involves photogrammetry techniques carried out with RPA, analysis of the digital images obtained through vector measurement in the QGIS software, and comparison using binarization and pixel recognition techniques in raster images. The study concludes that the methods employed proved effective for monitoring the constructive pathology (crack) in the dome of the Church of Saint Francis of Assisi, although they present limitations regarding the accuracy of the metrics obtained. The vector method was shown to be susceptible to errors resulting from manual contour marking, while the matrix method presented an inaccuracy of approximately two square centimeters, related to the pixel size used. It is noteworthy, however, that this latter method has potential to be automated in future applications.