Area-Wide estimation of seismic building structural types in rural areas by using decision tree and local knowledge in combination

Deep Multimodal Fusion Model for Building Structural Type Recognition Using Multisource Remote Sensing Images and Building-Related Knowledge

“…The main findings through experimental analysis are fourfolds: (1) Combining multi-source remote sensing images and BRK has great potential for recognizing BSTs. It take a step forward on the foundation of Geiß et al [1] and An et al [19].…”

“…However, since the BST is a description of the internal structure of a building, it is challenging to infer BSTs directly based on remote sensing images alone, which hinders the largescale application of these methods. Given a particular coupling law between BRK and building structure, other methods [1], [18], [10], [19] focus on combining BRK and remote sensing images to improve the reliability of BSTs recognition. These methods usually obtain BRK such as tax assessor and cadastral data [1], [18] from related government agencies and mapping software [10], [19].…”

“…Given a particular coupling law between BRK and building structure, other methods [1], [18], [10], [19] focus on combining BRK and remote sensing images to improve the reliability of BSTs recognition. These methods usually obtain BRK such as tax assessor and cadastral data [1], [18] from related government agencies and mapping software [10], [19]. They design handcraft features based on this knowledge and fuse them with the handcraft features of remote sensing images.…”

“…SAR has a high sensitivity to the backscattering characteristics of objects [40], which can reflect the roughness of the building surface and help distinguish the materials of the house roof and exterior walls. BRK, including roof type, color, and group pattern, is coupled with BSTs and can assist in recognizing BSTs [19]. By leveraging the combination of multi-source remote sensing images and BRK, a more comprehensive understanding of BSTs can be achieved.…”

Benefits of global earth observation missions for disaggregation of exposure data and earthquake loss modeling: evidence from Santiago de Chile

“…Taubenböck et al (2009), Borzi et al (2011), Qi et al (2017) characterize the built environment with remote sensing data and retrieve specific fragility functions or damage probability matrices, respectively. In contrast to that, e.g., Borfecchia et al (2010), Geiß et al (2015a, 2017b, 2018, Liuzzi et al (2019), Liu et al (2019), Torres et al (2019), and An et al (2021) combined limited in situ ground truth information characterizing the building inventory with features from remote sensing and use techniques of statistical inference for a complete labeling of the residual building inventory according to specific vulnerability levels or more generic properties such as construction material or occupancy, respectively. Related methodological principles were also exploited by, e.g., Wieland et al (2012Wieland et al ( , 2016, Wieland (2013), Geiß et al (2016), Pittore et al (2020), andFan et al (2021) to assess seismic vulnerability or related parameters on a coarser spatial level to allow for the use of data with larger spatial coverage.…”