A New Algorithm to Classify the Homogeneity of ERS-2 Wave Mode SAR Imagette

Abstract

A new classification parameter is developed using 1535 ERS-2 wave mode synthetic aperture radar (SAR) test imagettes to better differentiate homogeneous and inhomogeneous imagettes. The comparison between the new parameter (Min) and the previous one (Inhomo) (Schulz-Stellenfleth and Lehner, 2004) was done under varied threshold values of Inhomo. It is concluded that the performance of 'Min' is much better than 'Inhomo' when applying to the 1535 test imagettes. Furthermore, both Min and Inhomo are applied to nearly 1 million imagettes collected for the period from 1 September 1998 to 30 November 2000. The comparisons of the global inhomogeneous distribution between 'Min' and 'Inhomo' reveal that both the areas and percentage of inhomogeneity calculated by 'Min' are larger than that calculated by 'Inhomo'. By analyzing the low wind speed distribution of HOAPS data, we found that low wind speed over the ocean is one of the key reasons for the inhomogeneity of SAR imagettes.

A Case Study of Polar Low Detection Using ERS-2 Wave Mode Image

“…It may be noted that the algorithm was empirically developed mainly for the retrieval of integral ocean wave parameters from synthetic aperture radar data. It was later applied and referred by several studies including our earlier works [8]. The wind speeds retrieved using this algorithm from the three images shown in Fig.…”

“…ERS-2 images are quite useful for several studies. For example, our previous study uses them for validating and developing a new algorithm for their classification [8]. A number of SAR missions are crucial for continuous SAR research.…”

Multi‐Task and Multi‐Scale Intelligent Electromagnetic Sensing with Distributed Multi‐Frequency Reprogrammable Metasurfaces

“…By now, many RF sensing strategies have been proposed for this purpose from the detection of the notable body actions [ 7–11 ] to the recognition of subtle body movements. [ 12–16 ] However, they are hardly deployed for at‐home health care and beyond in practical sceneries, since either they are prohibitive due to the use of massive costly sensors, or suffer from the seriously limited temporal‐space resolution arising from the mechanical movement. Recently, reprogrammable metasurface (RM) has been introduced as a key enabling part of EM sensing systems for breaking the aforementioned limitations, [ 17–23 ] enabling the EM sensing to achieve many advanced functionalities that were thought to be impossible.…”