Remote sensing is all about collection of data about objects without any physical contact. It uses the solar electromagnetic (EM) spectrum of atmosphere its base energy. The electromagnetic spectrum is divided into visible region, infra red region (NIR), and microwave region of remote sensing. The remote sensing done in visible, NIR, Mid infra red and Thermal Infrared from 0.3 up to 15 micrometers is called optical remote sensing since its energy can be focused through optics. While the microwave region or energy is used by radar remote sensing sensor system. In optical remote sensing the system is basically composed of: the sun as the source of energy, the atmosphere as transmission path, the object needs to be either observed by or transmit through or reflected from and the satellite sensor. The solar radiation is transmitted through the atmosphere before reaching the surface. Thus reached radiation is reflected back to the sensor through the atmosphere. The radiation which crosses through the atmosphere involves in absorption and dispersion process. On the other hand in microwave remote sensing the source of energy is not the sun but the radar antenna in the satellite. The important advantage of microwave remote sensing is that due to longer wavelengths the effect of atmosphere is negligible where in optical remote sensing it will get scattered by some of the atmospheric scatterers e.g.O2, water vapor, CO2 and aerosol. Wavelength is an important characteristic of the EM, which determines the penetration capacity of the energy. The longer wavelength of radar remote sensing enables high penetration capability and very useful for measuring the bio-physical properties of vegetation and forest.
Many natural phenomena such as carbon dioxide concentration, hydrological cycle, energy balance in atmosphere are related to the climate and its change. But they are also related to Earth's vegetation cover such as forests. Forest is one of the most significant global vegetation cover in its area and volume. Forests are like the lungs of the earth. Forests help earth breathe by turning carbon dioxide into oxygen. . Therefore it is important to monitor and safeguard forests. With the advancement in the field of geo-information science remote sensing microwave remote sensing has been effectively used to detect and monitor forest parameters (e.g. area, height, density, volume, etc.). This is due to its capacity to monitor the environment independent of solar energy and weather conditions, especially in tropical and sub-tropical regions, where cloud cover are almost continuously present. The monitoring may be on the biophysical characteristics or the components such as leaves branches or trunk of the trees . The advantage of Polarimetric radar is ability to utilize single signal frequency with different polarization and there by helps in acquiring more information without the complexity in construction of radar. The information generated by this method is more when compared with conventional synthetic aperture radar (SAR) which operates in single, fixed-polarization for both transmitting and receiving the radio waves are more .
In order to preserve all the scattered information, polarimetric form of storing information is needed. In this the measured information are vector measurement which is stored in the form of scattering matrix. This scattering matrix can be decomposed into coherency or covariance matrix with process enabling amplitude and phase. The review of decomposition theorems were explained in . The classification of dominant scatterer was explained in .
The objective of this current study is improving the forward scattering model developed in  called as Freeman-II model. This model is developed to understand the geophysical parameter present in the polarimetric radar backscatter for air borne data set. The Freeman-II model contains two component scattering mechanism namely volume scattering which is modeled as randomly oriented prolate spheroid and double-bounce which represent ground-trunk interaction or surface scattering modeled by a pair of orthogonal surfaces with different dielectric constants, thus derived models are shown in methodology section. The output parameters from this model are the backscatter coefficient from each of the two components and two parameters which describing them. The backscatter contributions estimated from the model can be used to estimate the contribution of each HH, VV and HV backscatter terms and HH-VV phase difference. Without using any ground truth, Freeman-II model fits the two component scattering mechanism to the polarimetric SAR backscatter data. This model is justified as a simple model when compared with other models for forward scattering as developed in  ,  where model inputs are more than outputs from the model. This model was developed with the assumption reflection symmetry media which compensates for polarization orientation angle that yields covariance matrix reflection symmetry. This reflection symmetry condition also proved, in scattering matrix the coefficients correlating the co-polarized and cross-polarized terms are zero .
The polarization orientation angle is the angle of rotation about the line of sight. For horizontal medium these shifts are zero and for objects which are not horizontal to the line of sight, the orientation angle produces some shift. These shifts results in reducing cross-polarized intensity and producing covariance matrix reflection symmetry which is not correct. The compensation in orientation angle also results in decreasing the volume scattering power which leads to misinterpretation on the observed information . So these shifts are needs to be considered when modeling scatterer. The problem with Freeman-II model is which explicitly assumes reflection symmetry for canopy model. The specification of Freeman for further improvement also states the incorporation of orientation angle consideration in canopy scattering model .
The study conducted in  also states the problem of orientation angle and generalized it by adding new component and also modified the canopy model for different distribution of tree trunk and branches. This modification yields asymmetry matrices .Previous studies have revealed that Freeman-II model have been applied in airborne sensors ,  ,  but no research have been done for space borne data by considering the orientation angle shift.
This present study aims at incorporating the orientation angle shift to the model developed by Freeman to understand its effect in space borne Advanced Land Observation Satellite (ALOS) Phased Array L-band Synthetic Aperture Radar (PALSAR) data.  Such proposed model in rest of the proposal will be called as Modified-Freeman-II model (M-Freeman-II).
2 Research Identification
2.1 Research Objectives
Based on the work done so far the primary objective is:
To develop forward scattering model for multilayered vegetation of tropical forest area using polarimetric space borne SAR data and to understand the scattering mechanism of randomly oriented tree leaves, due to double bounces and due to stem.
Specific Objectives are:
To understand the influence of orientation angle shift for the improvement in identifying different scatterers.
To investigate the improvement in the result produced by the M-Freeman-II model compared with Freeman-II model.
2.2 Research questions
Which polarimetric SAR parameter is used to represent the polarization orientation shift of tree trunks and surface scatterer?
What could be the effect of Signal to noise (SNR) in identifying different scatterer?
What could be the improvement in the results produced by M-Freeman-II model?
How to validate the improvement in the result after modification in freeman-II model?
2.3 Improvement aimed at
Freeman-II model consist of two components to represent volume scatterer and double-bounce scatterer present in the polarimetric backscatter data. As per literature, the volume scattering power is affected by the orientation angle of polarization. This work aims at identifying the orientation shift and trying to incorporate the change for the application of Freeman-II model for space borne data set with modification incorporating the polarization orientation angle shift.
2.4 Related Work
Earlier modification in orientation angle shift was carried out as in  using covariance matrix which reported that the modification yields better results for airborne data.
The symmetry property of polarimetric scattering elements is discussed in  for the distributed targets that also includes symmetry relation of geophysical directional features.
The relation between various polarimetric backscatter coefficients were studied with symmetries under consideration such as reflection, rotation, azimuthal and centrical symmetry. It also explains about the correlation between co-polarized and cross polarized elements  .
The study in  investigates the effect of orientation angle compensation using coherency matrix. It utilizes the models in ,  for the investigation, and shows the importance of orientation angle for volume scatterer mechanism.
3 Project set-up
3.1 Method adopted
This research work aims at developing forward scattering model for different scatterer mechanism present in the backscatter image generated from the ALOS PALSAR data. The available data is of the raw format (i.e. Level 1). So the processing steps involved in generation of backscatter image from the polarimetric raw data are
Polarimetric Calibration: Needs to be done reduce the problems created by the fully polarimetric system.
Polarimetric Speckle Filtering: Speckle corrupt the polarimetric observables i.e. the phase and intensity. Specific methods have to be adopted to reduce the randomness in the signal acquired.
Methodology Flow Diagram
Polarization Synthesis: At this stage the formation of scattering matrix in any polarization will be done.
Polarimetric Features: It is used to represent the information present in the scattering matrix.
Orientation Angle shift estimation :The model referred from  for the estimation of orientation angle shift is related to azimuth slope and range slope by the following equation 
tan Î¸ =
Where, Î¸ is Orientation angle shift, is azimuth slope, slope in ground range direction, is radar look angle.
For the estimation of orientation angle from the coherency matrix the equation selected from 
Where, Î¸ = denotes real part and to limit the range of from -45â-¦ to + 45â-¦, to avoid rotating the matrix in wrong axis. , are the elements of basis set of the coherency matrix, values can be obtain from phase and magnitude of channels of polarimetric radar.
This model has been selected because its effectiveness has been verified with interferometric SAR techniques in  , . After the estimation of the orientation angle it should be compensated before applying the decomposition as per the literatures.
Polarimetric Decomposition: For understanding the Modification the decomposition will be done using both Freeman-II and M-Freeman-II scattering models.
For the volume scattering from the forest the Freeman-II model utilizes the randomly oriented prolate spheroids. By considering a set of set randomly oriented prolate spheroids with orientation according to uniform phase distribution, the covariance matrix to represent volume scattering was modeled.
The M-Freeman-II model will be utilizing the model developed for the estimation of orientation angle shift including azimuth slope and ground slope in . For double-bounce scattering corner reflector, different dielectric constants were used for backscatter from ground-trunk interaction and surface alone. The covariance matrix thus obtained consists of ð›¼ parameter to represent backscatter. To represent ð›¼ value for ground-trunk interaction and surface interaction equation with separate reflectance coefficients for vertical surface and horizontal surface were used.
The power image will be generated with contribution from different from different scatterers. Scattering mechanism contribution like (HHcan(db), HVcan(db),can(db), HHgd(db), VVgd(db), HH-VV, Phasegd(deg) will be estimated and plotted. These results have to be compared with the Freeman-II output by plotting them and its validity can be verified if it shows values of orientation shift under the canopy also.
In  after the assumption property of reflection symmetry , and which tells that co-polarized and cross-polarized term in the coefficients will be equal to zero. And the covariance matrix will be
With the above assumption which tells (ShhS*hv SvvS*hv 0) the component derived to explain the volume scattering using randomly oriented prolate spheroid.
The component of volume scatterer in Freeman-II model is:
Volume Scattering =
Double-bounce from ground-trunk interaction =
Where, is the correlation between HH and VV polarization, ð›¼ which represents double bounce (ground-trunk and surface return) is relative amplitude.
The fg, fc is the canopy and ground scatter contribution to the HH cross section. With separate equation for ð›¼ the phase value in double-bounce model for ground-trunk scatter and surface scatter can be discriminated.
Direct surface Return
Where, the vertical surface (Trunk) has reflection coefficient Rth and Rtv for horizontal and vertical polarization respectively. The horizontal surface (ground) has reflection coefficients Rgh and Rgv.The Î² represents attenuation coefficient. The Ï’ are real and they represent phase delay in radar pulse in horizontal and vertical polarization. The term h represents Height of the tree.
After ð›¼ is solved the value for fg, ,fc can be solved from total backscatter model
<MhhM*hh> = fg + fc ; <MhvM*hv> = fc ; <MhvM*hv> = fc + |ð›¼|2 fg ; <MhhM*vv> = fc + ð›¼fg
Freeman-II Polarimetric scattering model:
= fc volume + fg double bounce
3.2 Plan of the project
Output Expected at different Phases
Freeman-II model analysis
At this phase the research question one can be answered. And the identified parameter can be further analyzed for extension. The expected outcome at this phase is the identification of model that could be relevant for the improvement aimed at from literatures.
Understanding, Modification & developing M-Freeman-II model
At this phase the research question two and three would be answered. The expected output from this phase is the images with difference in values of backscatter. Improvement in identifying volume scatterer with increase in intensity for cross-polarization.
The outputs provided in the earlier phase are to be validated in order to address the research question four by plotting estimated scattering mechanism contributions from M-Freeman-II with Freeman-II model output and testing its validity by analyzing the power loss in double bounce is equal to power gain in volume scattering.
3.3 Risks and contingencies
This current study aims at improving the Freeman-II model, so expected problems may be software crashing since free source software will be used; interpretation problems of obtained results are possible. Thus this is planned to get more interaction with the supervisors. Alternative software is available to continue with the work in case the main polSARpro fails.
4 Resources required
To complete the project necessary data, software and good working environment is required.
The data proposed to be used in this project is ALOS PALSAR fully polarimetric data. Two scenes of same area are proposed to be used and information is provided in tables:
Table 1. Scene 1 information Table 2. Scene 2 information
4.4 Software and hardware
Software which will be used for project: PolSARpro, Nest and Beam.
Minimum specification of computer hardware: 512MB RAM 18 GB HD and 500MHz processor
5.1 Staff already consulted
Shashi Kumar, Scientist "C" (IIRS guide)
Dr.Yousif Hussin (ITC guide)
5.2 Communication plan
With IIRS supervisor meeting will be 2 hours weekly.
With ITC guide I will communicate through mail at the end of every week with my Reflection diary which will tell my daily work and for necessary help will as through mail itself.