PSI - Issue 64

H.F. Stewart et al. / Procedia Structural Integrity 64 (2024) 573–579 Stewart, Cusson, Greene Gondi & Oliver / Structural Integrity Procedia 00 (2024) 000–000

576

4

Fig. 3. Tri-stereo DSM (left) and multispectral image (right) illustrating the west abutment area of Samuel de Champlain Bridge. Black arrows indicate temporary structures on the central portion of the bridge deck.

Panchromatic-band (PAN) images at 50 cm native resolution from the tri-stereo image triplet dated 3 April 2021 were used to create a DSM at 1 m Ground Spacing Distance (GSD). The central image of the tri-stereo set was georeferenced using an existing permanent ground control monument in the Montreal metropolitan area (MERNQ, 2024), for which the coordinate values and elevations were acquired from the survey monument datasheet published by the Réseau Géodésique du Québec. The two remaining images of the tri-stereo triplet were then co-registered to the central image and orthorectified using the High-Resolution Canada Digital Elevation Model (HR-CDEM). Finally, the orthorectified tri-stereo image triplet was converted into a 1 m resolution DSM with projected coordinates in the Canadian Spatial Reference System v2 (NAD83 CSRS), Universal Transverse Mercator Zone 18N (UTM 18N), with heights relative to the Canadian Geodetic Vertical Datum of 2013 (CGVD 13). The DSM was edited in point cloud editing software to remove obvious gross noise (mostly around the bridge’s main span tower) and points on transient structures on the bridge deck. These structures were confirmed to be transient or temporary by reviewing panchromatic and multispectral images from the tri-stereo set and were typically associated with construction activity, such as vehicles or temporary road covering structures such as the ones shown in Fig. 3. A cleaned DSM point cloud was then created, clipped so that only the Samuel de Champlain Bridge and immediate surroundings were included. BlueMarble GlobalMapper V24 was used to convert the clipped DSM into WGS84 latitude, longitude, and elevation coordinates, and exported to GeoTiff format for use in SAR geocoding and height error estimation. 2.2. Satellite tri-stereo DSM and lidar DSM comparison To evaluate DSM vertical accuracy, the satellite tri-stereo DSM was compared with a historical airborne Lidar survey (DQ, 2015) acquired in 2015 before the construction of the SdC Bridge, which has horizontal and vertical accuracies of ±20 cm (ASPRS, 2023). Point cloud tiles were downloaded from the Canada Open Data Portal, projected into NAD 83 (CSRS) UTM 18N coordinates, reduced to CGVD 13 using the CGVD 28 – CGVD 13 separation of - 0.371 m at the survey control monument, and water elevations (≤10.6 m CGVD 13) removed. The resulting Lidar DSM was edited to remove points associated with temporary or moving features, such as ships at the Port of Montreal. Height accuracy was assessed by comparing Lidar and tri-stereo DSM heights at sample points on the old Champlain Bridge and the nearby Estacade ice control structure and Clément Bridge, as well as the Concorde Bridge, the portion of Jacques Cartier Bridge from the west shore of Sainte-Hélène Island to the east abutment, and the portion of Victoria Bridge east of Saint-Lambert Locks of the Saint Lawrence Seaway. The structures selected for comparison have flat road surfaces without steel trusses or superstructures, similar to the SdC Bridge approach spans (away from the main span tower). Points were sampled on clear, unobstructed areas of the roadway, away from road sign gantries, temporary structures, or construction equipment. In the case of the old Champlain Bridge, which was partially deconstructed at the time of tri-stereo image acquisition, points were selected on the remaining surface of the bridge deck. Vertical distances between Lidar and tri-stereo DSMs were compared at each point, and the height differences were used to compute Root Mean Square Error (RMSE) and linear vertical error at 90% confidence interval (LE90).