Master of Science
Electrical & Computer Engineering
Hyperspectral imaging (HSI) is a technology that captures three dimensions of information, two spatial dimensions and a spectral dimension. A focal plane, which is used to collect the intensity information of a scene, is only two-dimensional. In a conventional camera this is not a problem as it collects only two-dimensional images, but since hyperspectral imagers collect three-dimensional images this presents a problem. How do you collect these three-dimensional images? There have been clever schemes to use the focal plane to collect three-dimensional data as well as different capture methods which make either a spatial or spectral dimension the third dimension that is collected with subsequent scans. The next logical question is how to characterize and calibrate these systems. A radiance calibration is described and results in a low error calibration. For most systems, established methods will not work therefore, new metrics need to be developed and new merits created to fully measure and calibrate these systems. For line scan systems, a stationary measurement of a knife edge is not possible and movement is necessary for capturing the target. Two methods, stare step and continuous, were used and the results were similar to each other. In this paper, 3D noise methodology is considered for hyperspectral imagers (HSI), but there are not three dimensions to evaluate for an HSI as the spectral dimension cannot be used. Therefore, the concepts from 3D noise are used to discover that the spatial noise of the imager is the main source of noise in the sensor.
Dissertation or thesis originally submitted to ProQuest
Wade, Jonathon Michael, "Characterization of Hyperspectral Imagers" (2023). Electronic Theses and Dissertations. 2990.