Document Type

Thesis - Open Access

Award Date

2017

Degree Name

Master of Science (MS)

Department / School

Electrical Engineering and Computer Science

First Advisor

Morakot Kaewmanee

Second Advisor

Helder Dennis

Keywords

BRDF correction, drift est

Abstract

Given their low level of temporal, spatial, and spectral variability, Pseudo-Invariant Calibration Sites (PICS) have been increasingly desired as data sources for radiometric calibration of Earth imaging satellite sensors. The temporal resolution for PICS data acquired by any sensor is limited by the amount of time required for it to make subsequent passes over the site. Consequently, for any given PICS, it can take many years of imaging to develop a sufficient amount of cloud-free data to perform radiometric calibration; this can be especially problematic for sensors in their early years after launch. This thesis presents techniques to combine Landsat-8; normally acquiring data for every 16 days, image data from multiple PICS into a single dataset with increased temporal resolution and is called “PICS Normalization Process” or PNP. Landsat-8 Operational Land Imager (OLI) data from six Saharan desert sites were normalized to the Libya-4 reference. The normalized data were then merged into a “Super PICS” dataset, and the estimation of calibration drift was derived. The results of the Super PICS dataset show that the temporal resolution of the calibration dataset can be increased by approximately a factor of three to four times. The normalization process was performed on radiometrically and geometrically corrected image data (“L1T” product), and also on the same image data corrected for BRDF effects using a quadratic function of the solar zenith angle and TOA reflectance over a region of interest. An additional uncertainty analysis was performed using the BRDF corrected image data based on the following parameters which are involved in this whole BRDF PICS Normalization Process: Worst-case histogram bin analysis, Temporal Uncertainty of each PICS, BRDF Super PICS uncertainty. The resulting uncertainties are within the currently accepted satellite calibration range, within 3% for all spectral bands. Overall, the process indicates a calibration drift for OLI within 0.15% per year, agreeing quite well with the calibration drift derived from the on-board calibrators.

Library of Congress Subject Headings

Artificial satellites in remote sensing -- Calibration.
Landsat satellites -- Calibration.
Imaging systems -- Image quality.

Description

Includes bibliographical references (pages 81-83)

Format

application/pdf

Number of Pages

100

Publisher

South Dakota State University

Download

Share

COinS
 

Rights Statement

In Copyright