Towards Precision Agriculture for oil palm mineral nutrition management: relationships between the reflectance spectrum of oil palm leaves and nutrient deficiencies.
This study aims at determining if the reflectance spectra of oil palm leaves can provide an alternative measurement tool to expensive and time-consuming foliar chemical analysis to get information about any deficiencies of the trees. We focus here on the relationships between the reflectance spectra of oil-palm leaves and their deficiencies in five elements: nitrogen (N), phosphorus (P), potassium (K), magnesium (Mg), and iron (Fe). We carried out statistical analyses over a database that we built on palm trees measured in Sumatra (Indonesia), combining leaves reflectance in the visible and near-infrared domain (450-900 nm), leaves chemical composition, and observations of deficiencies visual symptoms on the tree. We obtained insignificant results on nutrient concentration prediction due to error rates above the stress detection threshold. At least we were able to discriminate the extreme situations of stress with an acceptable precision. We discuss here the results in terms of different sources of noise and errors in methodology, and we give prospective ways of improving the analysis.
| Main Authors: | , , , , , , |
|---|---|
| Format: | Texto biblioteca |
| Language: | ng |
| Published: |
Kuala Lumpur : MPOB,
2007
|
| Subjects: | Agricultura de precisión., Nutrición mineral., Palma de aceite, |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | This study aims at determining if the reflectance spectra of oil palm leaves can provide an alternative measurement tool to expensive and time-consuming foliar chemical analysis to get information about any deficiencies of the trees. We focus here on the relationships between the reflectance spectra of oil-palm leaves and their deficiencies in five elements: nitrogen (N), phosphorus (P), potassium (K), magnesium (Mg), and iron (Fe). We carried out statistical analyses over a database that we built on palm trees measured in Sumatra (Indonesia), combining leaves reflectance in the visible and near-infrared domain (450-900 nm), leaves chemical composition, and observations of deficiencies visual symptoms on the tree. We obtained insignificant results on nutrient concentration prediction due to error rates above the stress detection threshold. At least we were able to discriminate the extreme situations of stress with an acceptable precision. We discuss here the results in terms of different sources of noise and errors in methodology, and we give prospective ways of improving the analysis. |
|---|