Issue 46

F. Bazzucchi et alii, Frattura ed Integrità Strutturale, 46 (2018) 400-421; DOI: 10.3221/IGF-ESIS.46.37

To date, 73% of accuracy has been encountered, and after the second campaign no evolution from previous damage degree has been noticed. Because of the assembled arrangement, a new set of analyses has been carried out in classifying the damage degree of each precast element (6 for each span). A preliminary result has been shown in Fig. 25(b). To recognize each element, a manual segmentation on the image has been carried out, but as a future goal the system will be trained to autonomously recognize the bridge structural members. In the panorama of large-scale application, an interesting work has been recently published [43]. The pattern recognition is focused on detecting the modal shapes and free vibrations of a bridge by analyzing the huge amount of data stream of smartphones residents that crosses the bridge. In this case, many raw data and an intelligent DL algorithm could capture the desired dynamic features with acceptable accuracy.

Figure 25 : mismatch of damage degree (a) ; element-scale analysis (b) .

Figure 26 : Weight detection systems.

418