Adonis
Beyond wet or dry. Measure how much.
Fine-grained leaf wetness estimation from complementary radar signal views and calibration.
A useful representation preserves the physical quantity, not just a class label.
The question
A binary wet/dry label misses how surface water changes. Can radar measure degrees of leaf wetness?
The idea
Combine range–Doppler, range–phase and range–azimuth information with contrastive feature learning and inference-time calibration.
| Setting | Adonis | Traditional sensor |
|---|---|---|
| Controlled | 4.43 | 11.84 |
| Field | 6.49 | 14.32 |
What the evidence shows
The paper reports Leaf Wetness Level mean absolute error of 4.43 in controlled settings and 6.49 in field conditions. Traditional leaf wetness sensors report 11.84 and 14.32, respectively, in that evaluation. These are errors on the paper’s LWL scale, not percentages or claims across arbitrary crops and environments.
Source: IEEE INFOCOM 2025What it opens up
A physical representation must preserve the quantity that matters, with calibration under the conditions in which it will be used. This is a foundation for asking how such estimates should be maintained over time.