Imaging Science Thesis Defense: A New Vicarious Technique for Radiometric and Spatial Calibration of Drone-based Multispectral and Hyperspectral Imaging

Event Image
CIS Ph.D Defense
Imaging Science Thesis Defense
A New Vicarious Technique for Radiometric and Spatial Calibration of Drone-based Multispectral and Hyperspectral Imaging Systems

David Conran
Imaging Science
Rochester Institute of Technology

Register for Zoom Link Here

Abstract:
A new technology (convex mirrors) used for radiometric and spatial characterization of remote sensing imaging systems has demonstrated equal, but unique results when compared to traditional targets (Lambertian panels) for vicarious calibration. Using convex mirrors to characterize both the radiometric and spatial response of hyperspectral and multispectral imaging systems has yet to be thoroughly investigated for drone-based platforms. Convex mirrors produce radiometrically accurate point sources that can be imaged by remote sensing systems.
Assessing new techniques for vicarious characterization of imaging systems requires thorough comparison to traditional methods. The Empirical Line Method (ELM) employs the use of Lambertian panels to characterize an imaging systems response to varying surface reflectances for radiometric calibration. The slanted edge method for extracting the spatial response of imaging systems also uses Lambertian panels to form a high contrast edge. Spatial performance characterization defines the smallest spatial feature that can be observed and assists in the discovery of poorly focused or highly distorted optical configurations.
The research findings indicate that convex mirrors provide a novel technique to extract new information when deployed for drone-based imaging experiments. The reflected sunlight produces a broadband point source for characterizing the spatial misregistration of HSI systems without the need to oversample a traditional edge target. More importantly, the reflected sunlight from a convex mirror can be defined in a closed-form radiometric expression that can easily be used to validate the small target radiometric performance of imaging systems, as discussed in the research. Small target performance of HSI systems has drastic consequences for sub-pixel target detection and spectral unmixing, for example. Further discoveries unveiled the inconsistent radiometric performance of a well-calibrated HSI system using our point targets in the presence of platform motion and orthorectification compared to Lambertian panels.
   
Intended Audience:
All are Welcome!

To request an interpreter, please visit myaccess.rit.edu

Contact
Lori Hyde
Event Snapshot
When and Where
April 12, 2024
9:00 am - 11:00 am
Room/Location: Zoom
Who

This is an RIT Only Event

Interpreter Requested?

No

Topics
research