Noncontact Sensing Techniques for AI-Aided Structural Health Monitoring: A Systematic Review

Sabato, Alessandro; Dabetwar, Shweta; Kulkarni, Nitin Nagesh; Fortino, Giancarlo

doi:10.1109/jsen.2023.3240092

Cited by 38 publications

(20 citation statements)

References 163 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Heat leaking through a building's facades causes increasing economic losses as the building ages. As a result, rapid and advanced non-destructive evaluation (NDE) methods are constantly sought to detect the location of this heat loss, create effective maintenance plans, and increase the energy efficiency of the building envelope 1 . Heat can be lost through the building via doors, windows, and sub-surface damage, which are not detected when traditional visual inspection is performed.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Application of mask RCNN for localizing and quantifying areas of energy leak in buildings using infrared images

Dabetwar

Niezrecki

Sabato

2023

Health Monitoring of Structural and Biological Systems XVII

View full text Add to dashboard Cite

As buildings age, energy begins to leak through various locations such as window seals, walls, subsurface cracks, and damaged areas, even in seemingly healthy structures. Most of the time, areas of energy loss remain undetected because they are not visible to the naked eye. Due to the increasing amount of energy lost through such areas and defects which has an impact on overall energy efficiency. However, infrared images (IR) can be used to detect energy leaks as well as identify subsurface damages. Infrared thermography (IRT) is a popular method for assessing the condition of buildings and infrastructures. While IRT can provide information about the location and severity of energy leaks, manually analyzing the collected data can be a cumbersome process. As a result, there is a need to automate the detection of the areas from where energy is lost. Image segmentation methods based on deep learning algorithms can effectively automate the inspection process. In this study, an approach based on a pre-trained mask region-based convolutional neural network (mask RCNN) is proposed for the first time in conjunction with IR images to localize and quantify areas of heat loss. Mask RCNN demonstrated significant accuracy in identifying the location and quantifying the size of the area of heat loss in inspected buildings with above 99% confidence.

show abstract

Section: Introductionmentioning

confidence: 99%

“…New techniques based on transfer learning 14 , ResNet 15 , and AlexNet 16 are being implemented along with IR images. However, no attempts have been made in the literature to leverage DL along with IRT to automatically detect and quantify heat loss through buildings 1,17 .…”

Section: Introductionmentioning

confidence: 99%

Application of mask RCNN for localizing and quantifying areas of energy leak in buildings using infrared images

Dabetwar

Niezrecki

Sabato

2023

Health Monitoring of Structural and Biological Systems XVII

View full text Add to dashboard Cite

show abstract

“…Traditional contact-based measurements for structural health monitoring (SHM) come with some drawbacks, the most important one being the inability to provide full-field measurements. On the contrary, computer vision (CV) and photogrammetry methods inherently overcome this limitation and are gaining popularity in the SHM community 1,2 . Among the available photogrammetry techniques, three-dimensional (3D) digital image correlation (DIC) has been increasingly used in the SHM community to extract shapes, displacements, and strain profiles from pictures acquired using synchronized stereo cameras 3,4 .…”

Section: Introductionmentioning

confidence: 99%

Sensor-aided camera calibration for three dimensional digital image correlation measurements

Bottalico

Valente

Niezrecki

et al. 2023

Health Monitoring of Structural and Biological Systems XVII

View full text Add to dashboard Cite

Stereovision systems can extract full-field three-dimensional (3D) displacements of structures by processing the images collected with two synchronized cameras. To obtain accurate measurements, the cameras must be calibrated to account for lens distortion (i.e., intrinsic parameters) and compute the cameras’ relative position and orientation (i.e., extrinsic parameters). Traditionally, calibration is performed by taking photos of a calibration object (e.g., a checkerboard) with the two cameras. Because the calibration object must be similar in size to the targeted structure, measurements on large-scale structures are highly impractical. This research proposes a multi-sensor board with three inertial measurement units and a laser distance meter to compute the extrinsic parameters of a stereovision system and streamline the calibration procedure. In this paper, the performances of the proposed sensor-based calibration are compared with the accuracy of the traditional image-based calibration procedure. Laboratory experiments show that cameras calibrated with the multi-sensor board measure displacements with 95% accuracy compared to displacements obtained from cameras calibrated with the traditional procedure. The results of this study indicate that the sensor-based approach can increase the applicability of 3D digital image correlation measurements to large-scale structures while reducing the time and complexity of the calibration.

show abstract

“…Similarly, the use of machine learning (ML) for SHM has gained popularity due to robust, advanced, and easy-to-use algorithms [10][11][12] . Traditional autoencoders (AEs), a family of deep learning algorithms initially proposed to learn features of dynamics systems from datasets with reduced dimensions 13 , especially in the fluid mechanics and weather forecasting domains 14 , are becoming increasingly used for SHM 10,15 . Usually, these algorithms rely on the mean absolute error (MAE) or mean square error (MSE) as loss functions to reduce the reconstruction losses.…”

Section: Introductionmentioning

confidence: 99%

Time-inferred autoencoder for learning and adapting changes in the dynamic system

Kulkarni

Sabato

2023

Health Monitoring of Structural and Biological Systems XVII

View full text Add to dashboard Cite

Recent machine learning (ML) algorithms have resulted in new paradigms for extracting spatio-temporal characteristics (STCs), such as frequency spectra of critical infrastructures and performing structural health monitoring. However, the accuracy of the STCs extracted using ML is affected by any noise in the data used to train and test the ML algorithms. While noise reduction methods have been successfully proposed, they are application-specific, and none consider the dynamics of the targeted system. Hence, a novel framework named time-inferred autoencoder (TIA) is proposed. The TIA is based on a long, short-term memory (LSTM) neural network to learn the dynamics of the system and a maximum correntropy loss function for noise removal. The robustness of the TIA is validated by collecting a video of an undamaged beam for training the framework and learning the structure’s STCs. Later, the capability of the trained TIA to adapt to changes in the system’s dynamics and reconstruct the STCs is validated by recording noise-corrupted videos of a beam in three different damaged configurations. Results of laboratory tests showed that the TIA reconstructs the natural frequencies of the structure with an error of less than 1%. If further developed, the proposed framework can be used as a structural dynamics tool given its robustness and capability to adapt to changes and noise in the system.

show abstract

Noncontact Sensing Techniques for AI-Aided Structural Health Monitoring: A Systematic Review

Cited by 38 publications

References 163 publications

Application of mask RCNN for localizing and quantifying areas of energy leak in buildings using infrared images

Application of mask RCNN for localizing and quantifying areas of energy leak in buildings using infrared images

Sensor-aided camera calibration for three dimensional digital image correlation measurements

Time-inferred autoencoder for learning and adapting changes in the dynamic system

Contact Info

Product

Resources

About