Non-contact structural displacement measurement using Unmanned Aerial Vehicles and video-based systems. (November 2021)

• Record Type: Journal Article
• Authors: Ribeiro, D.; Santos, R.; Cabral, R.; Saramago, G.; Montenegro, P.; Carvalho, H.; Correia, J.; Calçada, R.
• Is Part Of: Mechanical systems and signal processing. Volume 160(2021)
• Publication Date: 2021-11
• Subjects: Civil Engineering structure -- Displacement -- Unmanned Aerial Vehicle -- Video system -- Inertial Measuring Unit -- Validation; Structural dynamics -- Periodicals; Vibration -- Periodicals; Constructions -- Dynamique -- Périodiques; Vibration -- Périodiques; Structural dynamics; Vibration; Periodicals; 621
• DOI: 10.1016/j.ymssp.2021.107869
• Languages: English
• ISSNs: 0888-3270

Abstract

Highlights: Methodology for estimating structural displacements based on UAV video systems. Efficient numerical integration strategy for IMU data processing through the application of high-pass digital filters and a rigorous control of distortion errors. Validation in the laboratory test showed an excellent agreement between the video and LVDT displacement time-histories, with a peak value error of 0.096 mm and an RMS error of 3.1%. Validation in the field-trail demonstrated a good agreement between the video and IMU displacement time-histories, with a peak value error of 1.47 mm and an RMS error of 9.3%. Abstract: This article describes an innovative methodology for estimating in-plane horizontal displacements of Civil Engineering structures based on video systems integrated on Unmanned Aerial Vehicles (UAVs). As the structure and the UAV are both in motion, estimating absolute structural displacements involves, first, assessing the relative displacements of the UAV-structure based on target tracking, and second, subtracting movement of the UAV based on the data derived from an embedded Inertial Measuring Unit (IMU). Computer-vision processing tools based on heuristic features were developed for target tracking, while an efficient numerical integration strategy was implemented for processing IMU data. A successful integration process requires a baseline correction of the records through the application of high-pass filters and a rigorous control of distortion errors. The validation of the proposed methodology was based on exploratory dynamic tests performed in the laboratory and in the field. The laboratory test involved measuring the displacements of a moving target, positioned over a seismic table using a Linear Variable Differential Transformer (LVDT), a DC accelerometer and the UAV video system in a stationary position. An excellent agreement was achieved for the three estimated displacements, both in the time and frequency domains. In particular, the comparison between the video and the LVDT displacement records showed a peak value error of 0.096 mm and a Root Mean Square (RMS) error of 3.1%. The field test consisted of measuring the displacements of a target fixed to a massive Reinforced Concrete (RC) wall based on the UAV video system in motion. Under these circumstances, the target does not experience any movement, and therefore, the virtual displacement of the target estimated by the video is due exclusively to the self-movements of the UAV. The comparison between the video and the IMU displacement records shows a good agreement, with a peak value error of 1.47 mm (15.5% relative error) and a RMS error of 9.3%. The results of this study are a step forward in estimating the absolute displacements of a vibrating structure based on camera-based measurement from UAVs, specifically with the support of IMU systems for the elimination of measurement inaccuracies as a result of the UAV motion.