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

575

3

Fig. 1. Saint Lawrence river and selected bridges in the Montreal metropolitan area (QC), Canada. Background image credit: ESRI base map and Airbus Pleiades image from 20 September 2021. Inset image credit: pontsamueldechamplain.ca. 2. Materials and methods 2.1. Satellite multispectral imagery registration and DSM creation An Airbus Pleiades tri-stereo image triplet dated 3 April 2021 was used for this study. The Pleiades constellation consists of two satellites (1A & 1B) operating in a sun-synchronous orbit at 694 km altitude, with a nominal 26-day repeat pass cycle (Astrium, 2012). Tri-stereo image sets include one scene acquired with the satellite approaching the area of interest, one scene acquired with the area of interest near the satellite nadir, and one scene acquired looking backward. Fig. 2 shows an example of stereo and tri-stereo image viewing geometry (ADS, 2021). Tri-stereo viewing geometry provides three unique image pairs for DSM generation while minimizing shadow zones seen in stereo pairs, which is useful in densely built urban areas. Acquiring the stereo image set on the same along-track pass also helps to minimize radiometric and atmospheric variations (Hobi and Ginzler, 2012). The absolute geolocation accuracy (including both horizontal and vertical accuracies) of Pleiades tri-stereo imagery is ±10.5 m and relative accuracy is ±1.5 m at the 90% confidence level (CE90). Comparisons of vertical error in tri-stereo DSMs have been shown to have vertical Root Mean Square Errors (RMSE) less than 3 m overall and less than 1.5 m on bare ground (Panagiotakis et al., 2018; Loghin et al., 2020).

Fig. 2. Stereo and tri-stereo satellite viewing geometry. Adapted from Airbus Defence & Space Intelligence.