DEEP INFRARED SENSING IN HYPERSONIC/SPACE ENVIRONMENTAL CONDITIONS
Abstract
Infrared (IR) sensing in the extreme environmental conditions will become more and more important for military applications due to the development of hypersonic/space flights. Only highly-limited measurable information could be used for IR sensing because of a significant change in the optical window, resulting from its exposure to the extreme environments. Towards target detection under extreme conditions, we first require to find a new concept of IR sensing method that can accurately measure an extremely hot object, e.g., > 2000 K, and to predict the optical properties of IR optical window to be exposed to the extremely high temperature. Here, we will develop a computational IR spectroscopy that enables the reconstruction of a target spectrum from pixel level signal responses through the desired spectral filters, using the computational information recovery based on the deep neural network. We also aim to develop the computational restoration of distorted IR images captured through the optically-changed optical window due to the hot temperature, which is theoretically based on a combination of a convolutional neural network and long short-term memory. The proposed work, i.e., realization of IR sensing platform for detecting targets at the extremely high temperature comprising deep learning-based IR spectroscopy and image restoration, can pave the way for efficient, adaptive, easy-to-apply IR sensing, which are highly useful for military purposes and many IR imaging applications.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Apr 20, 2023
- Source ID
- FA23862214027
Entities
People
- Yunsang Kwak
Organizations
- Air Force Office of Scientific Research
- Kumoh National Institute of Technology
- United States Air Force