19 Nov 2024

Deep Learning Approach for Accurate Segmentation of Sand Boils in Levee Systems

Abstract: Sand boils can contribute to the liquefaction of a portion of the levee, leading to levee failure. Accurately detecting and segmenting sand boils is crucial for effectively monitoring and maintaining levee systems. This paper presents SandBoilNet, a fully convolutional neural network with skip connections designed for accurate pixel-level classification or semantic segmentation of sand boils from images in levee systems. In this study, we explore the use of transfer learning for fast training and detecting sand boils through semantic segmentation. By utilizing a pretrained CNN model with ResNet50V2 architecture, our algorithm effectively leverages learned features for precise detection. We hypothesize that controlled feature extraction using a deeper pretrained CNN model can selectively generate the most relevant feature maps adapting to the domain, thereby improving performance. Experimental results demonstrate that SandBoilNet outperforms state-of-the-art semantic segmentation methods in accurately detecting sand boils, achieving a Balanced Accuracy (BA) of 85.52%, Macro F1-score (MaF1) of 73.12%, and an Intersection over Union (IoU) of 57.43% specifically for sand boils. This proposed approach represents a novel and effective solution for accurately detecting and segmenting sand boils from levee images toward automating the monitoring and maintenance of levee infrastructure.

19 Nov 2024

Widened Attention-Enhanced Atrous Convolutional Network for Efficient Embedded Vision Applications under Resource Constraints

Abstract: Onboard image analysis enables real-time autonomous capabilities for unmanned platforms including aerial, ground, and aquatic drones. Performing classification on embedded systems, rather than transmitting data, allows rapid perception and decision-making critical for time-sensitive applications such as search and rescue, hazardous environment exploration, and military operations. To fully capitalize on these systems’ potential, specialized deep learning solutions are needed that balance accuracy and computational efficiency for time-sensitive inference. This article introduces the widened attention-enhanced atrous convolution-based efficient network (WACEfNet), a new convolutional neural network designed specifically for real-time visual classification challenges using resource-constrained embedded devices. WACEfNet builds on EfficientNet and integrates innovative width-wise feature processing, atrous convolutions, and attention modules to improve representational power without excessive over-head. Extensive benchmarking confirms state-of-the-art performance from WACEfNet for aerial imaging applications while remaining suitable for embedded deployment. The improvements in accuracy and speed demonstrate the potential of customized deep learning advancements to unlock new capabilities for unmanned aerial vehicles and related embedded systems with tight size, weight, and power constraints. This research offers an optimized framework, combining widened residual learning and attention mechanisms, to meet the unique demands of high-fidelity real-time analytics across a variety of embedded perception paradigms.

14 Nov 2024

Application of Deep Learning for Segmenting Seepages in Levee Systems

Abstract: Seepage is a typical hydraulic factor that can initiate the breaching process in a levee system. If not identified and treated on time, seepages can be a severe problem for levees, weakening the levee structure and eventually leading to collapse. Therefore, it is essential always to be vigilant with regular monitoring procedures to identify seepages throughout these levee systems and perform adequate repairs to limit potential threats from unforeseen levee failures. This paper introduces a fully convolutional neural network to identify and segment seepage from the image in levee systems. To the best of our knowledge, this is the first work in this domain. Applying deep learning techniques for semantic segmentation tasks in real-world scenarios has its own challenges, especially the difficulty for models to effectively learn from complex backgrounds while focusing on simpler objects of interest. This challenge is particularly evident in the task of detecting seepages in levee systems, where the fault is relatively simple compared to the complex and varied background. We addressed this problem by introducing negative images and a controlled transfer learning approach for semantic segmentation for accurate seepage segmentation in levee systems.

5 Nov 2024

Innovations of Cellular Automata

Purpose: In the past several years, there has been a rather substantial uptick in the amount of research within the realm of cellular automata due to its ability to produce complex, self-organizing behavior from simplistic rulesets. The capability to produce this behavior is essential to understanding artificial life and intelligence. This uptick has resulted in numerous novel directions for experimentation within this computational playground. This work summarizes a few of the most impactful directions that have resulted from this research.

19 Sep 2024

The Forefront: A Review of ERDC Publications, Summer 2024

Abstract: As the main research and development organization for the US Army Corps of Engineers (USACE), the Engineer Research and Development Center (ERDC) helps solve our nation’s most challenging problems. With seven laboratories under the ERDC umbrella, ERDC expertise spans a wide range of disciplines. This issue of the Forefront highlights several ERDC reports from FY22, many of which were highly recognized and widely downloaded. The Forefront team was honored in FY23 to receive both the Information Technology Laboratory’s Communication Award and the ERDC Communication Award for our Summer 2022 issue of the Forefront. The Forefront team and the Information Science and Knowledge Management Branch (ISKM) as a whole are committed to staying current with best practices and exploring new techniques to communicate ERDC’s research excellence. Through quality publications, dynamic presentations, and ongoing training opportunities, ISKM strives not only to support ERDC but also to blaze a path to clear, concise, and engaging scientific communication products. Remember, if it ever takes you more than five minutes to find an answer, contact us. We are here to help!

18 Sep 2024

Using iThenticate for ERDC Publications: Avoiding and Addressing Unintentional Plagiarism

Abstract: The US Army Engineer Research and Development Center (ERDC) conducts world-class research that supports national endeavors and the Army mission. To demonstrate the reliability of ERDC’s research and to preserve ERDC’s reputation, it is critical that ERDC publications meet quality standards. This includes reviewing publications for potential copyright infringement, which adds another level of assurance to the quality and integrity of ERDC’s published works. Therefore, this report aims to explain the benefits and purpose behind implementing iThenticate, a powerful antiplagiarism tool, into the ERDC In-formation Technology Laboratory–Information Science and Knowledge Management (ISKM) Branch’s publication process and to present thorough guidance on using iThenticate effectively. To accomplish this, this document outlines the basics of copyright law, how to use iThenticate, and how to provide proper attributions for both text and images. With this information, ISKM editors will be able to better communicate to authors the results of iThenticate reviews and to propose solutions for any issues that iThenticate may highlight.

9 Sep 2024

Leveraging Artificial Intelligence and Machine Learning (AI/ML) for Levee Culvert Inspections in USACE Flood Control Systems (FCS)

Abstract: Levee inspections are essential in preventing flooding within populated regions. Risk assessments of structures are performed to identify potential failure modes to maintain the safety and health of the structure. The data collection and defect coding parts of the inspection process can be labor-intensive and time-consuming. The integration of machine learning (ML) and artificial intelligence (AI) techniques may increase accuracy of assessments and reduce time and cost. To develop a foundation for a fully autonomous inspection process, this research investigates methods to gather information for levees, structures, and culverts as well as methods to identify indicators of future failures using AI and ML techniques. Robotic plat-form and instrumentation options that can be used in the data collection process are also explored, and a platform-agnostic solution is proposed.

5 Sep 2024

From Research to Production: Lessons Learned and Best Practices

Abstract: This paper provides an overview of best practices to assist individuals and teams in transitioning software from a research product into a production environment. The information contained in this paper consists of best practices and lessons learned from an assignment consisting of transitioning a science-based research suite of programs into a more modern software format with appropriate preparations and considerations to be deployed in a production environment. The original software suite was written using both MATLAB and Python programming languages, and the new production version was written in the Python programming language.

4 Sep 2024

Comparison of Numerical Simulations of Heat-Induced Stress in Basalt

Abstract: Energy losses due to excessive noise and heat are primary liabilities in traditional mining processes. Some of the currently researched methods to improve these liabilities involve heating the rock to induce internal stress fractures that make it easier to extract or remove rock with traditional mining equipment. Physical experimentation has yielded useful data that have been applied to numerical simulations of the heating and fracturing of rock, and multiple such simulations have been developed in the commercial multiphysics simulator COMSOL. Since COMSOL is not widely available on DoD high-performance computers, the goal of this research is to develop methods of replicating simulations developed in COMSOL as simulations that run in Abaqus FEA, another commercial multiphysics simulator. In this work, a simulated basalt cylinder with a 25 mm radius and a 158 mm height is subjected to a surface heat flux approximating the effects of a laser beam applied to the top of the cylinder. Simulated stress distributions, displacements, and temperatures obtained from both simulators are compared. When comparable results were not obtained using both simulators, the differences in results were investigated using simplified versions of the simulation.

29 Aug 2024

Experimental Evaluation of Steel Beams with Mechanical Section Reduction Retrofitted with Fiber Polymers

Abstract: Steel elements working in a harsh environment can be exposed to corrosion that degrades their performance and threatens the integrity of the whole structure. Recent studies propose using carbon (CFRP) and basalt (BFRP) fiber–reinforced polymers to repair corroded steel cross sections; however, most of these studies have not explored many of the structural characteristics, including ductility. In this study, we conduct a series of full-scale experimental tests to investigate the impact of corrosion, represented as mechanical section reduction, on steel beams as well as the impact of repairing the beams using CFRP and BFRP in enhancing the beams’ structural performance. Mechanical section reduction, introduced to the flange and web elements, is used to establish a baseline dataset that captures the impact of repairs in the absence of corrosion. Four-point bending loading conditions are utilized for all tested beams. The results show that the reduction of the flange and web section lowers the beams’ yielding load by 10% and 8%, respectively, compared with a beam with a full cross section. Utilizing CFRP and BFRP patches can partially restore the corroded beams’ ductility; however, the BFRP is outperforming the CFRP in improving their ultimate strength by 10% and enhancing their ductility by 10%.