Underexposed Vision-Based Sensors’ Image Enhancement for Feature Identification in Close-Range Photogrammetry and Structural Health Monitoring

Abstract

This paper describes an alternative structural health monitoring (SHM) framework for low-light settings or dark environments using underexposed images from vision-based sensors based on the practical implementation of image enhancement algorithms. The proposed framework was validated by two experimental works monitored by two vision systems under ambient lights without assistance from additional lightings. The first experiment monitored six artificial templates attached to a sliding bar that was displaced by a standard one-inch steel block. The effect of image enhancement in the feature identification and bundle adjustment integrated into the close-range photogrammetry were evaluated. The second validation was from a seismic shake table test of a full-scale three-story building tested at E-Defense in Japan. Overall, this study demonstrated the efficiency and robustness of the proposed image enhancement framework in (i) modifying the original image characteristics so the feature identification algorithm is capable of accurately detecting, locating and registering the existing features on the object; (ii) integrating the identified features into the automatic bundle adjustment in the close-range photogrammetry process; and (iii) assessing the measurement of identified features in static and dynamic SHM, and in structural system identification, with high accuracy.

Unmanned Aerial Vehicle-Based Structural Health Monitoring and Computer Vision-Aided Procedure for Seismic Safety Measures of Linear Infrastructures

“…Depending on the field of view, the UAV should fly in front of the object, like in the example in Figure 1 , or above the object, as shown later in the validation test and pipeline shake-table tests. Similar to vision-based vibration SHM using steady cameras, as previously studied by the authors [ 35 , 36 , 37 , 38 , 39 ], the UAV camera should also be kept stable, and the UAV body should not drift while monitoring the tests; otherwise, they will affect the data accuracy. Therefore, this study proposes conversion and correction steps between two key steps for UAV-based seismic SHM, i.e., the computer vision-aided procedure and seismic safety measures, as shown in Figure 1 , with the details given in the next subsections.…”

“…The pixel distribution shifts more to brighter areas, while in darker areas, the pixel counts are slightly reduced. The quality of underexposed or low-contrast images is then improved using the Contrast-limited Adaptive Histogram Equalization (CLAHE) method [ 40 ], which was studied previously by the authors [ 39 ], along with several image enhancement algorithms to ensure that their impacts are insignificant to the data accuracies. Next, each image is processed continuously using region detection as well as feature detection, extraction, and matching algorithms.…”

Characteristics of Deformation and Stability of Ultra-Deep Pit in Plateau Alluvial–Lacustrine Gravel Strata

“…Figure 6 shows the configuration of monitoring items at the top of the pit walls. Close-range photogrammetry technology uses images obtained from close-distancetarget photography to determine the spatial positions of manual marking points [31][32][33][34]. A FUJIFILM-XT20 non-metric camera was used in the lab-scale geophysical model test to this end.…”

Novel tools for port environmental management systems

“…Associated with automatic image analysis systems, computer vision techniques seem an excellent complement according to previous experiences in other fields for detecting and recognition [Arribas et al 2011, Eskandari et al 2020, Storbeck and Daan 2001. Computer vision techniques have recently experienced a quick evolution, being implemented in a wide range of different applications with high efficiency and performance [Chen and Li 2021, Dong and Na 2021, Ngeljaratan and Moustafa 2021. Deep learning on convolutional neural networks is proven to achieve very high performance on computer vision tasks [Leonard 2019].…”

“…Computer vision techniques, associated with automatic image analysis systems, have proven to be excellent tools in other fields for detecting and recognizing objects [Arribas et al 2011, Eskandari et al 2020, Storbeck and Daan 2001. Computer vision techniques have undergone rapid evolution in recent years, demonstrating high efficiency and performance across a wide range of applications [Chen and Li 2021, Dong and Na 2021, Ngeljaratan and Moustafa 2021. Deep learning on convolutional neural networks has particularly shown exceptional performance in computer vision tasks [Leonard 2019].…”