Introduction to structure from motion and its applications in remote sensing
- Los Alamos National Laboratory
This talk discusses my experience at Los Alamos National Laboratories developing the Wide Area Persistent Surveillance (WAPS) system AngelFire and the problems with working with low resolution surface models. This experience provided a motivation to seek solutions to utilize the redundant WAPS imagery to build surface models of the urban environment. Structure from Motion (SfM) is a process that takes multiple view imagery and compute the 3D structure of a scene. We will walk through the basic algorithm and discuss areas for optimization. Military services and intelligence agencies face long-standing challenges with processing, exploiting, and disseminating ISR data. The problem is that too much data is being produced and not enough people to look at it and the problem is not going away. As a result of this data overload, we need to shift the way we think about data and find creative ways to use and present it so it can be easily digested by decision makers. SfM also provides a means for developing a data processing and organization architecture. Applications for various remote sensing applications will be discussed for motivation for why SfM and Multi-View Stereo rendering is an important area that needs to be continued to be developed.
- Research Organization:
- Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)
- Sponsoring Organization:
- USDOE
- DOE Contract Number:
- AC52-06NA25396
- OSTI ID:
- 1046544
- Report Number(s):
- LA-UR-11-00210; LA-UR-11-210; TRN: US201215%%505
- Country of Publication:
- United States
- Language:
- English
Similar Records
Understanding Modern Magnets through Conformal Mapping
Quantum Calisthenics: Gaussians, The Path Integral and Guided Numerical Approximations