Date of Award

2026

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Computer Science

Committee Chair

Tathagata Mukherjee

Committee Member

Aaron Kaulfus

Committee Member

Brian Freitag

Research Advisor

Tathagata Mukherjee

Subject(s)

Image processing--Digital techniques, Remote sensing--Data processing, CubeSats

Abstract

The PlanetScope SuperDove constellation consists of approximately 130 CubeSats. Radiometric drift among the satellites degrades time-series workflows. We formulate inter-sensor harmonization as a supervised regression problem and train a single model across all eight spectral bands using overlapping regions from cross- sensor image pairs. Controlled experiments on same-day image pairs show that a cubic spline model with 20 uniform knots and joint Ridge regularization achieves the lowest error among the evaluated per-pixel regression approaches. The model converges to a Kullback-Leibler divergence between 0.001 and 0.003, and no alternative architecture studied achieves a statistically significant improvement. Models that fit on each spectral band independently generally perform worse, indicating that cross-band coupling contributes to harmonization performance. The cubic spline model performs poorly on cross-region image pairs, indicating that the harmonization fit on one region does not transfer to another and that training data quality has a greater impact on harmonization performance than model architecture.

Share

COinS
 
 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.