Position and Volume Estimation of Atmospheric Nuclear Detonations from Video Reconstruction
Abstract
Recent work in digitizing films of foundational atmospheric nuclear detonations from the 1950s provides an opportunity to perform deeper analysis on these historical tests. This work leverages multi-view geometry and computer vision techniques to provide an automated means to perform three-dimensional analysis of the blasts for several points in time. The accomplishment of this requires careful alignment of the films in time, detection of features in the images, matching of features, and multi-view reconstruction. Sub-explosion features can be detected with a 67 percent hit rate and 22 percent false alarm rate. Hotspot features can be detected with a 71.95 percent hit rate, 86.03 percent precision and a 0.015 percent false positive rate. Detected hotspots are matched across 57-109 degree viewpoints with 76.63 percent average correct matching by defining their location relative to the center of the explosion, rotating them to the alternative viewpoint, and matching them collectively. When 3D reconstruction is applied to the hotspot matching it completes an automated process that has been used to create 168 3D point clouds with 31.6 points per reconstruction with each point having an accuracy of 0.62 meters with 0.35, 0.24, and 0.34 meters of accuracy in the x-, y- and z-direction respectively. As a demonstration of using the point clouds for analysis, volumes are estimated and shown to be consistent with radius-based models and in some cases improve on the level of uncertainty in the yield calculation.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 24, 2016
- Accession Number
- AD1053867
Entities
People
- Daniel T. Schmitt
Organizations
- Air Force Institute of Technology