CCSS: Spectral Filter Array For Multispectral Imaging

University Of Dayton, Dayton OH

Investigators

Abstract

Objective: The objective of this project is to enable a new paradigm of spectral filter array (SFA) based image sensors in the visible wavelength range. This multispectral imaging problem is formulated as the task of making finite spatial-spectral measurements on a hyperspectral image signal with minimal risks for aliasing. The proposed SFA design exploits the periodic response of multilayer dielectric interference filters that has a natural interpretation in the Fourier sampling theories, giving direct correspondences between the refractive index contrasts, layer thicknesses and the fidelity of the captured spectral image. Intellectual merit: The intellectual merit is the understanding of how the spectral filters in image sensors affect the spatial spectral fidelity of the multispectral images acquired by SFA based sensors. The proposed joint processing-hardware solution is capable of capturing considerably sharper image than other SFA-based sensors considered to date, since extra spectral samples are measured with minimal sacrifices to spatial resolution. The expected outcomes of the research include multilayer thin film filters with optimal spectral response, a high fidelity SFA sensor, and an advanced post-capture processing algorithm. Broader impacts: The broader impact is the development of SFA based multispectral sensor for applications that require portable or simple, inexpensive solutions. The fidelity of spectral images directly affects food and water safety, remote sensing, geology, surveillance, biomedical imaging, environment monitoring, forensics, and mineralogy. The proposed educational and outreach program leverages various outlets to target local industries, a large graduate population comprised of practitioners of sensor exploitation, and tutorial sessions at high impact conferences.