Science Service System

Summary of Proposal LAN0054

TitleModelling of crop paramerers from microwave data
Investigator Dabrowska - Zielinska, Katarzyna - Institute of Geodesy and Cartography, Remote Sensing
Team Members
Prof dr Dabrowska-Zielinska, Katarzyna (KDZ) - Institute of Geodesy and Cartography, Remote Sensing
dr Lewinski, Stanislaw (LS) - Institute of Geodesy and Cartography, Remote Sensing
Budzyriska, Maria - Institute of Geodesy and Cartography, Remote Sensing
SummaryThe objectives of the project are: the elaboration of the efficient method of crop discrimination using Terra SAR-X data and Terra - SAR with ENVISAT ASAR data of various polarisation and angle; the development of new models which through inversion will give the canopy characteristics as various crop parameters which will give possibilities to assess crop growth and yield estimation. Extensive field measurements of various crop-soil parameters and crop recognition will be carried out simultaneously with satellites overpasses for the verification of the model deliverables. Synergy of various wave lengths and polarisation will be investigated in the context of crop and soil monitoring. The next approach will be to establish the relationship between the model parameters and soil- vegetation features providing the simulation of backscattering coefficient (using X and C multipolarization) and compare the model results with experimental data. For this approach the canopy cloud model of Atema and Ulaby will be applied. The inversion of the model against field measurements will allow estimating the canopy biophysical properties. It is obvious that during the season of crop growth the soil moisture various and therefore the backscatter will be modeled for different soil moisture and different crop what it will allow to interpret the backscatter value and understand its variability and distribution. Extending the canopy cloud model of Atema and Ulaby considered for the canopy as the single layer to one consisting more than one layer (as MIMICS) and combine optical and microwave data for soil moisture retrieval (under vegetated conditions) and replace the soil –sub model with an IEM to better reflect the interaction between vegetation layer and soil surface. Methods of classification - supervised as well as unsupervised methods will be considered. Classification will be performed for signatures calculated for fields as a whole or segments of fields. Neural network classifier based on multi-layer perceptron will be used for classification of images. eCognition method of object classification will be also applied. The test site is situated in the western part of Poland where since several years the measurement campaign has been carried out. The most common crops at the test site are cereals: wheat, rye, barley and triticale, which is a hybrid of wheat and rye. Rape, sugar beets, maize and alfalfa are also well represented. Data requirements: For the purpose of the project the Dual Polarisation Stripmap product radiometricaly enhanced will be used. Preferable acquisition starting the 1 of April until October. The elevation beam identification - STRIP 9 and STRIP 14, every 11 days with different dual polarisation: VV/HH; HV/VV respectively. Deliverables: classification maps, classification accuracy, results of modelling - various crop parameters The funding will covered by national project

Back to list of proposals

© DLR 2004-2016