← Back to Research Notes

Thermal remote sensing

A thermal anomaly is not yet an industrial event

A methodological note on baselines, mixed pixels, weather context, and why interpretation should remain separate from detection.

LAND SURFACE TEMPERATURETEMPORAL BASELINEMIXED PIXELSUNCERTAINTYVALIDATION
A thermal anomaly is not yet an industrial event
SAMPLE CONTENT

Illustrative sample writing and imagery; not completed research results.

Detection is a measurement decision

A thermal anomaly appears only after a reference condition, threshold, and comparison period have been chosen. Changing those choices can change which pixels are flagged.

One hot pixel can contain several surfaces

Roofs, pavement, bare ground, water, and atmospheric effects can contribute to the same observation. Spatial context matters before a source is named.

Weather belongs in the baseline

Air temperature, wind, humidity, and recent rainfall can shift surface temperature. A useful baseline should make those conditions visible rather than treating them as background noise.

Report candidates before causes

A cautious public layer can report repeated candidate patterns, observation support, and uncertainty without publishing a definitive claim about an individual facility.

View all notes →