Abstract's details
Level-2 assessment of along-track antenna pattern compensation for SAR altimetry
CoAuthors
Event: 2015 Ocean Surface Topography Science Team Meeting
Session: Instrument Processing: Measurement and retracking (SAR and LRM)
Presentation type: Type Oral
Contribution: PDF file
Abstract:
One of the main benefits of the along-track processing in SAR altimetry [1] is the speckle reduction that is achieved by multi-looking the single look echoes that are gathered for a given position on the Earth surface. Further advantages in terms of speckle reduction can be achieved on sea waveforms by compensating the along-track antenna pattern on the stacks of single look echoes before multi-looking.
This abstract is mainly aimed at investigating the precision of the geophysical parameters that are retrieved from the multilooked waveforms after the compensation of the Along-track antenna Pattern Compensation (APC).
Experiments have been conducted on CryoSat-2 acquisitions processed using the ESA GPOD service (SARvatore) to produce a wide dataset of multilooked waveforms with APC.
It is worth recalling here the standard processing chain for a SAR altimeter [1]: firstly an approximately equally spaced set of ground locations on the Earth surface, i.e. surface sample, is identified. A surface sample gathers a stack of single look echoes coming from the processed bursts during the time of visibility.
However, in a surface sample stack, each single look echo has been acquired from a different position of the instrument along the orbit and, as a consequence, it results to be scaled in power by the antenna pattern as function of the look angle. According to [2], on a uniformly rough spherical surface, e.g. the ocean, the power of the single look echoes in the stack is modulated by the along-track antenna pattern.
But the speckle being a multiplicative noise, the speckle noise power will also be shaped by the antenna pattern so that the speckle has not the same power in different single look echoes in the stack. Thus, by simply averaging the single look echoes, the speckle will not reduce as much as possible.
An increased speckle reduction can be obtained by compensating the power modulation due to the along-track antenna pattern on the stack before averaging. This way, the power of the speckle from the off-nadir beams is increased at the same power of the speckle in the central beams, so that the speckle in all the single look echoes in the stack is raised approximately to the same level. It is worth underlining here that, to correctly compensate the along-track antenna pattern, an accurate knowledge of the pitch is needed and the approach described in [2] has been here used, since it is well suited for ocean acquisitions. Moreover, together with APC the most off-nadir single looks echoes in the stack are ruled out from the multi-looking because they are expected to carry less information due to the limited receiving window of the CryoSat instrument and because their leading edge is expected to be much less steep than ones closer to the nadir.
Experiments have been conducted on CryoSat-2 acquisitions over ocean, aimed at verifying the improvements in speckle reduction and in precision of the retrieved physical parameters using APC in Level1 processing. The test dataset is planned to be composed by CryoSat acquisitions in SAR mode over both the Wadden Sea and open ocean, in order to test APC performance under a wide range of sea states.
As a metric to evaluate the speckle reduction, the Equivalent Number of Looks (ENL) has been used, which is defined as the estimate of the effective number of statistically independent looks and it is expected to be bounded by the number of averaged single look echoes. At first the experimental ENL has been evaluated on the two CryoSat Level1 datasets (without APC and with APC) as function of the SWH. Preliminary results show that an increase of about the 10% of average ENL along the waveforms is obtained using APC, independently of the SWH.
Moreover, we aim at verifying the possible improvements on the retrieved physical parameters. The SAMOSA retracker has been applied on the two CryoSat Level1 datasets (in case of APC an isotropical along-track antenna pattern was used in the retracker). Preliminary results show that that the misfit (here defined as the percentage root mean square error between the SAMOSA power waveform model and the 20Hz power waveforms) for the APC dataset is about 4% lower for SWH higher than 1 m that than for the dataset without APC applied. Further analysis have to be performed to assess the precision and the accuracy of the APC dataset in the sense of SSH, SWH and Sigma0.
[1] R. Raney, “The delay/doppler radar altimeter,” Geoscience and Remote Sensing, IEEE Transactions on, vol. 36, no. 5, pp. 1578–1588, Sep 1998.
[2] M. Scagliola, M. Fornari, and N. Tagliani, “Pitch estimation for cryosat by analysis of stacks of single-look echoes,” IEEE Geosci. Remote Sens. Letters, vol. PP, no. 99, pp. 1–5, 2015.
This abstract is mainly aimed at investigating the precision of the geophysical parameters that are retrieved from the multilooked waveforms after the compensation of the Along-track antenna Pattern Compensation (APC).
Experiments have been conducted on CryoSat-2 acquisitions processed using the ESA GPOD service (SARvatore) to produce a wide dataset of multilooked waveforms with APC.
It is worth recalling here the standard processing chain for a SAR altimeter [1]: firstly an approximately equally spaced set of ground locations on the Earth surface, i.e. surface sample, is identified. A surface sample gathers a stack of single look echoes coming from the processed bursts during the time of visibility.
However, in a surface sample stack, each single look echo has been acquired from a different position of the instrument along the orbit and, as a consequence, it results to be scaled in power by the antenna pattern as function of the look angle. According to [2], on a uniformly rough spherical surface, e.g. the ocean, the power of the single look echoes in the stack is modulated by the along-track antenna pattern.
But the speckle being a multiplicative noise, the speckle noise power will also be shaped by the antenna pattern so that the speckle has not the same power in different single look echoes in the stack. Thus, by simply averaging the single look echoes, the speckle will not reduce as much as possible.
An increased speckle reduction can be obtained by compensating the power modulation due to the along-track antenna pattern on the stack before averaging. This way, the power of the speckle from the off-nadir beams is increased at the same power of the speckle in the central beams, so that the speckle in all the single look echoes in the stack is raised approximately to the same level. It is worth underlining here that, to correctly compensate the along-track antenna pattern, an accurate knowledge of the pitch is needed and the approach described in [2] has been here used, since it is well suited for ocean acquisitions. Moreover, together with APC the most off-nadir single looks echoes in the stack are ruled out from the multi-looking because they are expected to carry less information due to the limited receiving window of the CryoSat instrument and because their leading edge is expected to be much less steep than ones closer to the nadir.
Experiments have been conducted on CryoSat-2 acquisitions over ocean, aimed at verifying the improvements in speckle reduction and in precision of the retrieved physical parameters using APC in Level1 processing. The test dataset is planned to be composed by CryoSat acquisitions in SAR mode over both the Wadden Sea and open ocean, in order to test APC performance under a wide range of sea states.
As a metric to evaluate the speckle reduction, the Equivalent Number of Looks (ENL) has been used, which is defined as the estimate of the effective number of statistically independent looks and it is expected to be bounded by the number of averaged single look echoes. At first the experimental ENL has been evaluated on the two CryoSat Level1 datasets (without APC and with APC) as function of the SWH. Preliminary results show that an increase of about the 10% of average ENL along the waveforms is obtained using APC, independently of the SWH.
Moreover, we aim at verifying the possible improvements on the retrieved physical parameters. The SAMOSA retracker has been applied on the two CryoSat Level1 datasets (in case of APC an isotropical along-track antenna pattern was used in the retracker). Preliminary results show that that the misfit (here defined as the percentage root mean square error between the SAMOSA power waveform model and the 20Hz power waveforms) for the APC dataset is about 4% lower for SWH higher than 1 m that than for the dataset without APC applied. Further analysis have to be performed to assess the precision and the accuracy of the APC dataset in the sense of SSH, SWH and Sigma0.
[1] R. Raney, “The delay/doppler radar altimeter,” Geoscience and Remote Sensing, IEEE Transactions on, vol. 36, no. 5, pp. 1578–1588, Sep 1998.
[2] M. Scagliola, M. Fornari, and N. Tagliani, “Pitch estimation for cryosat by analysis of stacks of single-look echoes,” IEEE Geosci. Remote Sens. Letters, vol. PP, no. 99, pp. 1–5, 2015.