Anàlisi comparativa d'algorismes operacionals d'estimació de paràmetres biofísics de la coberta vegetal amb teledetecció
- Autores: Aleixandre Verger i Ten
- Directores de la Tesis: Joaquín Meliá Miralles (dir. tes.), Fernando Camacho de Coca (codir. tes.)
- Lectura: En la Universitat de València ( España ) en 2008
- Idioma: español
- Tribunal Calificador de la Tesis: Vicente Caselles Miralles (presid.), María Amparo Gilabert Navarro (secret.), Frédéric Baret (voc.), José González Piqueras (voc.), José Luis Casanova Roque (voc.)
- Materias:
- Enlaces
- Tesis en acceso abierto en: TESEO
- Resumen
In the past ten years, various medium resolution sensors have been launched and many methods have been proposed to estimate biophysical vegetation parameters (FVC, LAI and FAPAR) from remotely sensed imagery in an operational way. To fully exploit the potential of current Earth observation programs and take advantage of the multiplicity of availble products, efforts have to be directed towards improving their consistency and accuracy. This needs validation and inter-comparison studies. Parallel development of new strategies for fusion of sensor measurements and derived products is also required. In this context, the main objectives of this thesis are:
1.- Evaluate in a comparative way the performances of different operational remote sensing approaches (LSA SAF, POLDER, VGT4AFRICA, GLOBCARBON, CYCLOPES and MODIS) for estimating FVC, LAI and FAPAR. And assess the discrepancies between different estimates.
2.- Explore the synergy of multi-sensor remote sensing signals and multi-algorithm biophysical retrievals for improving existing vegetation products. The performance of neural network based approach for estimating LAI from existing CYCLOPES/VEGETATION and MODIS products is assessed.
The research performed in this thesis is clearly in line with the present and future activities concerning the land surface monitoring. The first part of this thesis contributes to answer the strong requirement expressed by the users for a more comprehensive assessment of the quality and uncertainty of retrieval algorithms and vegetation products. The second part of this thesis investigates an innovative procedure to define algorithms able to assess vegetation properties with consistency and accuracy from multiple existing products and input measurements independently from data sources. This original investigation opens avenues for developing innovative sensor-independent algorithms in order to ensure the future service continuity avoiding gaps in products due to one sensor failure.
