Enabling seamless 3D semantic reconstruction from heterogeneous data at scale
Abstract
Human training and education in a simulated environment require itsfoundational 3D content to be geo-specific, accurate, semantically rich and timely. Tocreate such data, the current practice often needs to leverage a mixed use of data frommultiple sources which vary in resolution, scale and views, while this is yet laborintensiveand the process is normally ad-hoc to the type of datasets. This objective of thisresearch is to further automate this process by developing a set of enabling tools thatallow rapid registration, fusion and meshing 3D heterogeneous data, such that datacollected under constrained conditions can be rapidly, seamlessly processed to completeand high-fidelity 3D models. The proposed algorithms are expected to have a high degreeof flexibility and scalable to wide-area, such that processing data collected underdifferent means does not require high level of expertise and the produced data can beused to generate wide-area and high-volume integrated terrain dataset. These proposedtools, if successful, can greatly reduce the effort and lower the bars for on-demand 3Ddata collection for 3D content generation in support of simulated battlefield training,global situational awareness, and virtual terrain exploration.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Feb 17, 2020
- Source ID
- N000142012141
Entities
People
- Rongjun Qin
Organizations
- Office of Naval Research
- Ohio State University
- United States Navy