Cross-​modal feature based structural deformation monitoring using colored point clouds

Terrestrial laser scanning (TLS) is widely used in geodetic monitoring (including structural deformation monitoring) because of its direct 3D measurement capability and independence from natural light sources or supplementary illumination during scanning. However, methods rely only on point clouds may perform poorly when scanned surfaces have repetitive or even no distinct geometric features, where RGB images can complement to some degree. The fusion of point clouds and RGB images is underutilized, especially in structural deformation monitoring. Therefore, this research aims to investigate how structural deformation monitoring can benefit from image-assisted TLS point cloud processing using both geometric and radiometric information. The research will be divided into four work packages: alignment quality assessment of TLS point clouds and camera images (WP1), feature correspondence establishment based on geometry, Lidar intensity and RGB information (WP2), joint processing of high-dimensional features on specific structural deformation monitoring cases (WP3), and point cloud augmentation via high-resolution images (WP4). The findings of this research can be applied directly to structural deformation monitoring, as well as to other fields (e.g., virtual reality).