Navigating the accuracy-efficiency trade-off in dam surface deformation monitoring via terrestrial lidar and photogrammetric point cloud fusion

Abstract

A trade-off model is proposed that investigates the sensitivity of various point cloud fusion parameters, such as TLS density of stations, image redundancy in photogrammetry, voxel resolution, and ICP threshold, in relation to total operation time. A registration pipeline is proposed for point cloud fusion of TLS and photogrammetric point cloud data on the curved surfaces of dams. The proposed framework is validated for two periods of operation on a concrete gravity dam. The optimal configuration results in 2.8 mm RMS C2C and 4.6 mm LoD95. The proposed framework is capable of achieving 91% accuracy compared to TLS only and reducing operation time by 38%. The proposed framework is capable of providing guidelines for various scenarios of operation.