23 Apr 2020

PUBLICATION NOTICE: Use of Convolutional Neural Networks for Semantic Image Segmentation Across Different Computing Systems

ABSTRACT: The advent of powerful computing platforms coupled with deep learning architectures have resulted in novel approaches to tackle many traditional computer vision problems in order to automate the interpretation of large and complex geospatial data. Such tasks are particularly important as data are widely available and UAS are increasingly being used. This document presents a workflow that leverages the use of CNNs and GPUs to automate pixel-wise segmentation of UAS imagery for faster image processing. GPU-based computing and parallelization is explored on multi-core GPUs to reduce development time, mitigate the need for extensive model training, and facilitate exploitation of mission critical information. VGG-16 model training times are compared among different systems (single, virtual, multi-GPUs) to investigate each platform’s capabilities. CNN results show a precision accuracy of 88% when applied to ground truth data. Coupling the VGG-16 model with GPU-accelerated processing and parallelizing across multiple GPUs decreases model training time while preserving accuracy. This signifies that GPU memory and cores available within a system are critical components in terms of preprocessing and processing speed. This workflow can be leveraged for future segmentation efforts, serve as a baseline to benchmark future CNN, and efficiently support critical image processing tasks for the Military.

23 Apr 2020

PUBLICATION NOTICE: Optimized Low Size, Weight, Power and Cost (SWaP-C) Payload for Mapping Interiors and Subterranean on an Unmanned Ground Vehicle

ABSTRACT: Section 3 of the FY15 Force 2025 Maneuvers Annual Report indicates that in Dense Urban Areas (DUA), specifically in a subsurface, surface, or super-surface structure, the ability to identify threats will be diminished. Most commercially available LIght Detection And Ranging (LIDAR) systems are specifically designed for high-resolution aerial imaging and mapping applications. As a result, they tend to be large, heavy, power-hungry, data bandwidth intensive, and expensive. They also employ lasers that are not typically eye-safe, which limits their overall effectiveness in subterranean and the interiors of subsurface or super-surface structures. However, due to recent advances in the automotive industry, there are new generations of Size, Weight, Power, and Cost (SWaP-C) sensors that are eye-safe, making them suitable for use indoors and in subterranean environments. While these tradeoffs limit their effective use to hundreds of meters (compared to kilometers for their more expensive counterparts), they are ideal candidates for use in subterranean and building interiors. While cameras fill this niche to some extent, the volumetric calculations provided by these sensors provide additional intelligence to shape the security of the environment and offer more precision when maneuvering troops. These sensors would provide the warfighter with situational understanding in previously inaccessible locations. Therefore, to aid in the Army’s need to obtain and maintain situational understanding in DUAs, the authors propose utilizing low size, weight, power, and cost (SWaP-C) sensors, on a robot platform, for surveying and mapping underground structures and building interiors. Rapid/near real-time data processing is possible by utilizing open-source software and commercial off the shelf (COTS) components. Using the preferred sensor payload autonomously was also explored.

22 Apr 2020

Corps to resume Natomas levee work, close Garden Highway on April 23

The U.S. Army Corps of Engineers will resume major levee improvement construction in the Lower Natomas Basin beginning April 23, 2020, closing a portion of Garden Highway to all traffic.

22 Apr 2020

PUBLICATION NOTICE: The Forefront : A Review of ERDC Publications, Spring 2020

Abstract: The Engineer Research and Development Center (ERDC) is the premier civil works engineering and environmental sciences research and development arm of the U.S. Army Corps of Engineers (USACE). As such, it partners with the Army, Department of Defense (DoD), federal agencies, and civilian organizations to help solve our Nation’s most challenging problems in civil and military engineering, geospatial sciences, water resources, and environmental sciences. A special government knowledge center, ERDC Information Technology Laboratory’s Information Science and Knowledge Management (ISKM) Branch is critical to ERDC’s mission, fulfilling research requirements by offering a variety of editing and library services to advance the creation, dissemination, and curation of ERDC and USACE research knowledge. Serving as the publishing authority for the ERDC, ISKM publishes all ERDC technical publications to the Digital Repository Knowledge Core, sends a copy to the Defense Technical Information Center (DTIC) and creates a press release about each publication on the ERDC website. The Forefront seeks to provide an additional mechanism for highlighting some of our technical publications to the ERDC, USACE, Army, and DoD communities. This publication also encourages those outside ERDC to contact us about using ERDC editing services. For more information regarding the reports highlighted in this publications or others that ERDC researchers’ have created, please contact the ISKM virtual reference desk at erdclibrary@ask-a-librarian.info or visit the ISKM’s online repository, Knowledge Core, at https://erdc-library.erdc.dren.mil/ .