PSI - Issue 64

1762 Francesco Calabrò et al. / Procedia Structural Integrity 64 (2024) 1759–1766 Francesco Calabrò, Giovanna E. Minniti, Antonino Fotia, Raffaele Pucinotti / Structural Integrity Procedia 00 (2019) 000 – 000 One of the most important aspects, when it comes to interventions relating to the historical building heritage, is the structural component. Without prejudice to the subsequent in-depth information, specific to the executive project, it is necessary to be aware of the level of degradation of the load-bearing structures from the beginning, as the cost of their consolidation, in certain conditions, can become the main cost item. For this purpose, the most advanced structural monitoring techniques come in handy, capable of providing valuable information right from the start for the purposes of identifying the type of consolidation needed and, consequently, its cost impact. 3. Structural Health Monitoring Structural elements are prone to degradation and to developing cracks and lesions, with varying degrees of danger (Smith, J. 2018). Lesions can result from physiological responses or from structural damages necessitating monitoring for potential developments. Analyzing the degradation and the crack pattern and geometry is crucial for understanding it nature, causes, and severity. The examination involves assessing whether the issue is widespread or localized, interpreting the trend, form, and location of degradation and cracks on structural elements. Moreover, studying lesion or degradation progression over time is essential for understanding their evolution characteristics. If no progression occurs, the system has stabilized; otherwise, additional movements, crack increments, and deformations may ensue in search of a new equilibrium configuration (Garcia et al., 2015). The developing advanced survey and monitoring techniques for degradation and crack control by an automated monitoring system, is presented below. The methodology integrates advanced technologies, including automated drone flight (Liu, 2021), 3D modeling, high-resolution image processing, CNN-based deterioration detection (Convolutional Neural Networks), and structural analysis. It therefore allows not only the execution of the survey and the representation of the state of degradation and the crack pattern, but also and above all, to evaluate its variations over time. This is fundamental for identifying the type of consolidation necessary and, consequently, its impact in terms of cost. Moreover, it consents also important to better define the intervention priority. In fact, the urgency of an intervention is different in the case of structural failures still underway or already stabilized structural failures. The proposed methodology involves the design of drone flight plans and the automatic acquisition of images. The system is developed in 4 main phases namely (see Fig. 1): 1) Preliminary phase; including: (i) inspection and identification of the areas to be examined; 2) Flight phase; including: (i) Design of the flight plan and data acquisition; (ii) Flight and image acquisition; 3) Processing phase; comprising: (i) 3D model reconstruction and high resolution orthophoto and orthomosaic reconstruction; (ii) Deterioration detection using CNN; (iii) Correlation between deteriorated areas and structural elements; (iv) Visualization of deterioration on an orthomosaic model; (v) Structural analysis of the infrastructure for different degradation scenarios; 4) Visualization of results, including: (i) Creation of a historical database of points of interest; (ii) Querying the layer overlay database. The objective of using drones (with cameras with the same characteristics) and planning flight plans is to allow to always acquire images from the same position and with the same angle; these are necessary conditions to avoid pre processing for the images rectification. The objective of machine learning, however, is to extract information and correlations from the acquired data (following processing). In general, the planning of the flight plan and the determination of the image acquisition points require an initial inspection; this allows you to avoid interference with any obstacles during the various phases. Once the trajectory and the areas to be "acquired" within the individual frames have been established, using a flight planner, we move on to designing the trajectories and image acquisition points. This allows for the repeatability of inspection activities, making the acquired images easily comparable (in terms of increase in the state of degradation and/or damage, etc.).