Article
Multidisciplinary Sciences
Fanglin Bao, Xueji Wang, Shree Hari Sureshbabu, Gautam Sreekumar, Liping Yang, Vaneet Aggarwal, Vishnu N. N. Boddeti, Zubin Jacob
Summary: Machine perception uses advanced sensors to collect information for situational awareness. State-of-the-art machine perception faces difficulties with increasing number of intelligent agents. Exploiting omnipresent heat signal could be a new frontier for scalable perception. The proposed heat-assisted detection and ranging (HADAR) overcomes the challenge of ghosting and shows promising results compared to AI-enhanced thermal sensing.
Article
Construction & Building Technology
Shuai Teng, Zongchao Liu, Xiaoda Li
Summary: Automatic bridge surface defect detection is an important and efficient method for saving human resources and improving work efficiency. This study investigates the performance of different object detection algorithms in bridge surface defect detection, and proposes an improved YOLOv3 network as a decent detector for fast and real-time detection of bridge defects.
Article
Environmental Sciences
Carli J. Morgan, Matthew Powers, Bogdan M. Strimbu
Summary: Traditional inventories can be resource-intensive and require a trained workforce, but the use of handheld LiDAR and SfM algorithms show potential for efficient tree detection and measurement of dimensions and characteristics, such as defects and damages.
Article
Chemistry, Multidisciplinary
Fujun Du, Shuangjian Jiao, Kaili Chu
Summary: This paper proposes a lightweight target detection algorithm based on the YOLO algorithm, which enhances the accuracy of bridge structure damage target detection by using the BIFPN network structure for multi-scale feature fusion and the EFL for optimizing sample imbalance processing mechanism.
APPLIED SCIENCES-BASEL
(2022)
Article
Computer Science, Information Systems
Zhiyan Zhong, Hongxin Wang, Dan Xiang
Summary: Surface defect detection is critical in improving production yield of magnetic tiles. Existing methods struggle to accurately locate and segment small defects due to their low proportions and difficult visual identification. To overcome these challenges, we propose an effective algorithm that decomposes the image, estimates possible defect regions, locates defects, enhances contrast, and precisely segments defect areas. Experimental results demonstrate the algorithm's superiority and potential for industrial applications.
Article
Chemistry, Analytical
Yue Zhao, Bolun Chen, Bushi Liu, Cuiying Yu, Ling Wang, Shanshan Wang
Summary: This paper proposes an improved bearing defect detection algorithm based on YOLOv5, which includes gamma transformation and ResC2Net model. It achieves better performance in detecting and representing bearing defects through image preprocessing and feature extraction stages.
Article
Chemistry, Multidisciplinary
Wenguang Wang, Xiyuan Chang, Jihuang Yang, Gaofei Xu
Summary: This paper studies the LiDAR-based pedestrian detection and tracking with high-resolution sensing capability, which plays an important role in real-world applications. By using real-measured LiDAR point clouds, the problem of dense pedestrian detection and tracking is addressed, and a kernel density estimation method is used for center estimation and segmentation. Three novel features are defined for further performance improvement, and a new track management strategy is presented to deal with tracking instability caused by occlusions. The proposed method is validated on the KITTI dataset and shows significant improvements compared to existing methods.
APPLIED SCIENCES-BASEL
(2022)
Article
Construction & Building Technology
Mingzhu Wang, Srinath Shiv Kumar, Jack C. P. Cheng
Summary: This study proposes a framework based on deep learning defect detection and metric learning for tracking multiple sewer defects in CCTV videos. Experimental results show that tracking performance can be influenced by detection accuracy and configurations of the metric learning module.
AUTOMATION IN CONSTRUCTION
(2021)
Article
Engineering, Civil
Pravee Kruachottikul, Nagul Cooharojananone, Gridsada Phanomchoeng, Thira Chavarnakul, Kittikul Kovitanggoon, Donnaphat Trakulwaranont
Summary: This study developed a deep learning-based visual defect inspection system for reinforced concrete bridge substructures, which consists of four main components: image acquisition, defect detection, defect classification, and severity prediction. The system achieved high accuracy rates for defect detection, classification, and severity prediction, which were well-received and deemed suitable for practical application by Thailand's Department of Highways.
JOURNAL OF CIVIL STRUCTURAL HEALTH MONITORING
(2021)
Article
Optics
Yasser M. Fouda
Summary: The study introduces a new approach to fabric defect detection using integral images, which enhances computational efficiency and robustness. Experimental results demonstrate the efficiency and effectiveness of the proposed method in real-time detection of flawed fabric.
OPTICS AND LASER TECHNOLOGY
(2022)
Article
Chemistry, Multidisciplinary
Ichiro Ario, Yuta Hama, Khongkham Chanthamanivong, Yuki Chikahiro, Akimasa Fujiwara, Haicheng Ma
Summary: Large-scale natural disasters are occurring more frequently due to climate and environmental changes, resulting in increased vulnerability of infrastructure, especially bridges. Rebuilding damaged bridges and reconnecting transportation systems after disasters are major challenges. This study proposes an emergency bridge concept based on origami-inspired post-buckling theory and optimal deployable structure of scissors-type bridges to address these issues.
APPLIED SCIENCES-BASEL
(2022)
Article
Instruments & Instrumentation
Yongqiang Chen, Kai Luo, Liang Chen, Haobo Weng, Wei Liang
Summary: This study aims to accurately identify delamination defects in CFRP materials through the use of a hexagonal structure-based method. Experimental results demonstrate that this approach outperforms traditional square structures in terms of accuracy and efficiency in edge detection.
SMART MATERIALS AND STRUCTURES
(2022)
Article
Environmental Sciences
Xianjian Jin, Hang Yang, Xiongkui He, Guohua Liu, Zeyuan Yan, Qikang Wang
Summary: This paper proposes a novel LiDAR-based robust vehicle detection method, including point cloud clustering, bounding box fitting, and point cloud recognition. The method improves clustering quality, adaptability of bounding box fitting, and stability of fitting results. It also utilizes deep learning technique to achieve vehicle recognition. Experimental results demonstrate stable vehicle target recognition with a good balance between time consumption and accuracy.
Article
Engineering, Multidisciplinary
Dongjie Li, Zilei Zhang, Baogang Wang, Chunmei Yang, Liwei Deng
Summary: By improving the feature fusion module, loss functions, and model parameters of the YOLOX target detection algorithm, this article enhances the accuracy and speed of wood surface defect detection. Experimental results demonstrate significant improvements and advantages of the proposed method in detecting four types of defects on rubber timber.
Article
Environmental Sciences
Jaz Stoddart, Danilo Roberti Alves de Almeida, Carlos Alberto Silva, Eric Bastos Gorgens, Michael Keller, Ruben Valbuena
Summary: Current LiDAR-based methods for detecting forest change lack a biological link with ecosystem characteristics. We propose a model based on ecosystem morphological traits (EMTs) to predict forest aboveground biomass (AGB) change. By using a multitemporal dataset and selecting sensitive LiDAR metrics as proxies for each EMT, we can identify logging losses and regions of regrowth. The accuracy of our AGB estimation model is comparable to a statistically optimized model. Adoption of this EMT-based approach can improve the transferability and comparability of LiDAR models for AGB worldwide.