Device, Algorithm and Integrated Modeling Research for Performance-Drive Multi-Modal Optical Sensors
Abstract
Basic research was conducted in technology associated with adaptive, performance-driven multi-modal optical sensing, as well as the phenomenology of hyperspectral imaging of humans. Contributions included (1) An end-to-end simulation model demonstrated multimodal adaptive sensing in the context of vehicle tracking in an urban environment. (2) A feasible design was developed for a MEMS single-pixel tunable Fabry-Perot spectrometer using novel thermal activation. (3) A novel analytical model was developed for polarimetric imaging systems and used in an adaptive target detection algorithm to optimize polarizer angles. (4) A multimodal target-tracking algorithm was developed combining spectral and polarimetric imagery. (5) A novel technique for spectral waveband selection was developed and used in a vehicle tracking demonstration. (6) A database was developed of human spectral reflectance, which found that VNIR measurements of clothing were more robust than human skin or hair in distinguishing among pedestrians. Future work is recommended in developing the tunable FP device, rotatable polarizers, and algorithms for performance-driven sensing.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 17, 2012
- Accession Number
- ADA579303
Entities
People
- Alan D. Raisanen
- Andrew C. Rice
- Annette O. Rivas
- Jared A. Herweg
- John Kerekes
- Juan R. Vasquez
- Lingfei Meng
- Michael D. Presnar
- Sabino M. Gadaleta
- Tingfang Zhang
- Zoran Ninkov
Organizations
- Rochester Institute of Technology