This paper describes an image-based methodology for detection of structural cracks in concrete. The conventional approach based on line enhancement filtering has problems associated with inaccurate extraction of edge pixels. We therefore propose an edge-based crack detection technique consisting of five steps: crack width transform, aspect ratio filtering, crack region search, hole filling, and relative thresholding. In the first step, crack width transform, opposing edges are identified and classified as crack candidate pixels and a width map is generated. In the second step, aspect ratio filtering, the width map generated is used to remove noise. In the next two steps, the method searches for and restores missing pixels. In the last step, relative thresholding, residual noise is removed through reclassification of crack regions based on image-adaptive thresholding. The performance of this technique was tested using synthetic and real images. The results show that the proposed technique has greater accuracy and consistency in measuring crack width compared with the conventional technique.

The effects of illumination and shooting distance on crack image recognition were investigated by examining cracks in images taken with a camera. In order to examine the effects, images of cracks in a concrete structure taken while varying the illumination and shooting distance in an outdoor environment were analyzed. The images were acquired at a daytime illumination of 52,000 lx and a night illumination of 13 lx. The crack specimen images produced for the experiment were taken by increasing the shooting distance from 5 m to 100 m in each illumination. On the basis of the analysis on the modulation transfer function (MTF) and contrast sensitivity of the crack images, the effects of illumination and shooting distance on the sharpness of the crack images were investigated. The minimum crack widths that can be identified under each illumination were analyzed using MTF10 and Weber contrast 0.1, respectively. It was found that as the shooting distance increases, the effects of illumination on crack recognition become greater.

When diagnosing the condition of a structure, it is necessary to measure the widths of any existing cracks in the structure. To ensure safety when relying on images of cracks, the selected imaging parameters and processing technology must be well understood. In this study, the effects of the ground sample distance and threshold values on the crack width measurement error are analyzed from a theoretical perspective. Here, the main source of such errors is assumed to be due to the mixed pixel phenomena in the left and right boundary pixels. Thus, a mathematical model is proposed in which the intensity changes in these pixels are computed via an equation. In addition, the relationship between the error and error probability distribution is represented with an equation based on the threshold values and mean error. Upon analysis, it was found that the threshold value that minimizes the error is at the mid-point between the background and foreground, and the probabilistic nature of the error indicates that it is theoretically possible to predict both the error and its probability distribution. The proposed model was validated using artificial images.

Custom inspection using X-ray imaging is a very promising application of modern pattern recognition technology. However, the lack of data or renewal of tariff items makes the application of such technology difficult. In this paper, we present a data augmentation technique based on a new image-to-image translation method to deal with these difficulties. Unlike the conventional methods that convert a semantic label image into a realistic image, the proposed method takes a texture map with a special modification as an additional input of a generative adversarial network to reproduce domain-specific characteristics, such as background clutter or sensor-specific noise patterns. The proposed method was validated by applying it to backscatter X-ray (BSX) vehicle data augmentation. The Fréchet inception distance (FID) of the result indicates the visual quality of the translated image was significantly improved from the baseline when the texture parameters were used. Additionally, in terms of data augmentation, the experimental results of classification, segmentation, and detection show that the use of the translated image data, along with the real data consistently, improved the performance of the trained models. Our findings show that detailed depiction of the texture in translated images is crucial for data augmentation. Considering the comparatively few studies that have examined custom inspections of container scale goods, such as cars, we believe that this study will facilitate research on the automation of container screening, and the security of aviation and ports.

The structural deformation of a bogie frame manufactured using a composite material was monitored in real time using a distributed optical fiber sensor. The bogie frame contained an internally embedded standard single-mode optical fiber. Performance tests were conducted by applying a vertical load to the middle of the side beams on each side of the composite bogie frame. The strain distribution was monitored using an optical fiber sensor. A distributed optical fiber sensor system based on the Brillouin optical correlation domain analysis (BOCDA) technique with a 3 cm spatial resolution was used. The distributed strain measured using the optical fiber correlated well with the finite element (FE) analysis data, confirming that the composite bogie frame was fabricated as designed. For a vertical load of 182 kN, the maximum strain, which occurred in the middle of the side frame, increased by 1.3 times, as compared with a vertical load of 140 kN. The experiment was able to verify the balance and the structural stability of the left- and right-hand-side beams. Furthermore, it could confirm that there was a concentrated load where the side beam and the crossbeam meet, owing to a mismatch during the assembly of the composite bogie frame.