DeepSpecN: A new hybrid method combining PROSPECT-PRO and Conv-Transformer to estimate leaf nitrogen content from leaf reflectance
Shuai Yang, Anirudh Belwalkar, Dong Li, Yufeng Ge, Tao Cheng, Fei Wu, Longkang Peng, Daoliang Li, Kang Yu · 2025
Read DeepSpecN: A new hybrid method combining PROSPECT-PRO and Conv-Transformer to estimate leaf nitrogen content from leaf reflectance by Shuai Yang, Anirudh Belwalkar, Dong Li, Yufeng Ge, Tao Cheng, Fei Wu, Longkang Peng, Daoliang Li, Kang Yu (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13109343.txt.
Description
Shuai Yang and eight colleagues estimate leaf nitrogen from reflected light by wedding a physics model to a transformer network, in a 2025 paper. Nitrogen content tracks crop health, but field measurement is slow and destructive. Purely learned methods hunger for labeled leaves; purely physical ones wobble on inversion.
Their hybrid, DeepSpecN, trains on simulated spectra from a leaf optics model, so no field labels are needed. A wavelet step trims glare, an improved transformer learns the features, and a similarity screen picks the most realistic simulated samples, fighting domain shift. On over seventeen hundred maize leaves it beat rival methods, and it generalized to four other crops. A side finding: chlorophyll-based formulas predicted nitrogen better than protein-based ones. The authors offer the method for precision farming scouts who need chemistry from a camera.
Catalog record
| CREATOR | Shuai Yang, Anirudh Belwalkar, Dong Li, Yufeng Ge, Tao Cheng, Fei Wu, Longkang Peng, Daoliang Li, Kang Yu |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 63 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |