University of Twente Student Theses


Biophysical parameters retrieval for pearl millet using semi-empirical model

Ramathilagam, Arun Balaji (2020) Biophysical parameters retrieval for pearl millet using semi-empirical model.

Full text not available from this repository.

Full Text Status:Access to this publication is restricted
Abstract:The Polarimetric Synthetic Aperture Radar (PolSAR) uses an active microwave sensor which focuses on transmitting and receiving completely polarized microwaves for object detection and characterization. It exploits the amplitude and phase characteristics of the backscattering signals to characterize the targets. The standard of the Polarimetric Synthetic Aperture Radar (PolSAR) imagery, decompositions, and further analysis depends on the quality and the accuracy of its measurements. The polarimetric distortions in the observations lead to deterioration of the quality. These distortions are incurred due to non-ideal system behaviours and propagation errors, such as channel imbalance, crosstalk, and Faraday rotation. The polarimetric distortions are to be removed for accurate analysis using PolSAR data. In recent years compact polarimetric SAR has gained significant interest in scientific community as it operates with reduced hardware complexity, less data transmission rate, doubled swath width without compromising the resolution compared to a quad pol system. However, the calibration of hybrid-pol data is different as it measures signals in only two channels, unlike quad-pol data that uses redundancy in the measurements to estimate the calibration parameters. Moreover, most of the hybrid PolSAR decomposition and calibration algorithms assume a perfect circular transmit state. In reality, the transmit state cannot be perfectly circular due to system complexities. The study focuses on the calibration of hybrid PolSAR data for removing polarimetric distortions and to characterize the imperfect transmitter state of the hybrid pol sensor. Indian Space Research organisation’s Radar Imaging Satellite-1 (RISAT-1) hybrid polarimetric data over forest regions around Haridwar-Rishikesh valley in Uttarakhand state was opted for this study. Most of the hybrid PolSAR calibration algorithms use external calibration targets such as corner reflectors and Polarimetric active radar calibrators. However, calibration using the scattering properties of natural distributed targets (DTs) can reduce the required number of corner reflectors reducing the operational cost. The calibration of the data in this research is performed using forest regions that satisfy the properties of homogenous distributed targets and a single corner reflector. Hybrid pol decomposition modelling and Maximum normalized error (MNE) are used for testing the efficiency of the calibration technique. A significant improvement in the scattering mechanisms and target characterization is observed through decomposition modelling before and after calibration of the dataset. Subsequently MNE metric revealed that there is a decrease in the amount of distortions in the data before and after the calibration procedure.
Item Type:Essay (Master)
Faculty:ITC: Faculty of Geo-information Science and Earth Observation
Programme:Geoinformation Science and Earth Observation MSc (75014)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page