Calibration to maximize temporal radiometric repeatability of airborne hyperspectral imaging data

Area of Science:

• Remote Sensing
• Optical Physics
• Environmental Monitoring

Background:

• Temporal radiometric repeatability is a critical, yet understudied, aspect of airborne remote sensing data calibration.
• Existing calibration methods often do not adequately address variations introduced by multiple data acquisition times.

Purpose of the Study:

• To evaluate the temporal radiometric repeatability of airborne hyperspectral optical sensing data.
• To compare the effectiveness of different radiometric calibration methods under varying temporal conditions.

Main Methods:

• Acquired airborne hyperspectral data over 52 flight missions across three days using experimental objects.
• Applied four radiometric calibration methods: no calibration, empirical line method (ELM), and two atmospheric radiative transfer model (ARTM) approaches (ARTM and ARTM+).

Main Results:

• Spectral bands from 900-970 nm exhibited lower temporal radiometric repeatability than bands from 416-900 nm.
• ELM calibration was sensitive to flight time, sun parameters, and weather conditions.
• ARTM+ calibration demonstrated superior performance, significantly improving radiometric repeatability, particularly in spectral bands beyond 900 nm.

Conclusions:

• Airborne remote sensing data acquired across multiple time points can expect at least 5% radiometric error.
• Effective calibration, particularly ARTM+, is essential for reliable classification functions using temporally diverse data.
• Temporal replication in data acquisition is vital for accurately capturing variations and noise in remote sensing studies.