Automatic Road Tunnel Crack Inspection Based on Crack Area Sensing and Multiscale Semantic Segmentation

The detection of crack defects on the walls of road tunnels is a crucial step in the process of ensuring travel safetyand performing routine tunnel maintenance. The automatic and accurate detection of cracks on the surface of roadtunnels is the key to improving the maintenance efficiency of road tunnels. Machine vision technology combinedwith a deep neural network model is an effective means to realize the localization and identification of crackdefects on the surface of road tunnels.We propose a complete set of automatic inspection methods for identifyingcracks on the walls of road tunnels as a solution to the problem of difficulty in identifying cracks during manualmaintenance. First, a set of equipment applied to the real-time acquisition of high-definition images of walls inroad tunnels is designed. Images of walls in road tunnels are acquired based on the designed equipment, whereimages containing crack defects are manually identified and selected. Subsequently, the training and validationsets used to construct the crack inspection model are obtained based on the acquired images, whereas the regionscontaining cracks and the pixels of the cracks are finely labeled. After that, a crack area sensing module is designedbased on the proposed you only look once version 7 model combined with coordinate attention mechanism (CAYOLOV7) network to locate the crack regions in the road tunnel surface images. Only subimages containingcracks are acquired and sent to the multiscale semantic segmentation module for extraction of the pixels to whichthe cracks belong based on the DeepLab V3+ network. The precision and recall of the crack region localizationon the surface of a road tunnel based on our proposed method are 82.4% and 93.8%, respectively. Moreover, themean intersection over union (MIoU) and pixel accuracy (PA) values for achieving pixel-level detection accuracyare 76.84% and 78.29%, respectively. The experimental results on the dataset show that our proposed two-stagedetection method outperforms other state-of-the-art models in crack region localization and detection. Based onour proposedmethod, the images captured on the surface of a road tunnel can complete crack detection at a speed often frames/second, and the detection accuracy can reach 0.25 mm, which meets the requirements for maintenanceof an actual project. The designed CA-YOLO V7 network enables precise localization of the area to which a crackbelongs in images acquired under different environmental and lighting conditions in road tunnels. The improvedDeepLab V3+ network based on lightweighting is able to extract crack morphology in a given region more quicklywhile maintaining segmentation accuracy. The established model combines defect localization and segmentationmodels for the first time, realizing pixel-level defect localization and extraction on the surface of road tunnelsin complex environments, and is capable of determining the actual size of cracks based on the physical coordinatesystemafter camera calibration. The trainedmodelhas highaccuracy andcanbe extendedandapplied to embeddedcomputing devices for the assessment and repair of damaged areas in different types of road tunnels.

Early landslide mapping with slope units division and multi-scale objectbased image analysis——A case study in the Xianshui River basin of Sichuan,China

Previous studies on optical remote sensing mapping of landslides mainly focused on new landslides that have occurred, but little attention was paid to the early landslide due to its high concealment. In SAR technology, a prevalent method to detect early landslides, only can be used to identify the potential hazards of slow deformation. Therefore, it is necessary to explore new method of early landslides mapping by integrating all types of direct and indirect early features, such as cracks on slopes, small collapses inside and topographic features. In this study, an object-oriented image analysis method based on slope unit division and multi-scale segmentation was proposed to obtain accurate location and boundary extraction of early landslides. In the middle-and small-scale segmentation, the object, texture, spectrum, geometric features,topographic features, and other features were obtained to determine the local feature location of early landslides. The slope unit boundary was combined with the feature of a large-scale segmentation object to determine the scope of landslides. This method was tested in the Xianshui River basin in the Daofu County, Sichuan Province, China. The results demonstrate that:(1) Such features as landslide cracks and the small collapse at the bottom of slope can effectively determine the landslide position.(2) The slope unit division and the correct setting of shape factors in multiple segmentation can effectively determine the landslide boundary.(3) The accuracy of landslide location extraction was 83.33%, and the accuracy of boundary extraction for early landslides that were completely identified was evaluated as 82.67%. It is indicated that this method can improve the accuracy of boundary extraction and meet the requirements of the early landslides mapping.

Multiscale Hessian filter-based segmentation and quantification method for photoacoustic microangiography

The appearanee of blood vessels is an important biomarker to distinguish diseased from healthy tissues in several fields of medical applications. Photoacoustie microangiography has the advantage of directly visualizing blood vessel networks within mierocireulatory tissue. Usually these images are interpreted qualitatively. However, a quantitative analysis is needed to better describe the characteristics of the blood vessels. This Letter addresses this problem by leveraging an efficient multiscale Hessian filter-based segmentation method, and four measure- ment parameters are acquired. The feasibility of our approach is demonstrated on experimental data and we expect the proposed method to be beneficial for several microcireulatory disease studies.