POLARIZED POINT SCATTERERS: AN ALGORITHM FOR DETECTION USING ALOS-PALSAR DATA

This paper deals with a new class of localized point targets we call polarized point scatterers (PPS). They can be detected from Quadpol SAR data by using a two-dimensional filter, employing an eigenvalue/alpha threshold. They can then be used to estimate the dielectric constant of the target using a simple scattering model. We first consider the basic performance of the filter based on selection of thresholds and data window size to keep false alarms to a desired low level. We then illustrate application of the filter to sample data sets from the ALOS-PALSAR archive, showing its important application to urban areas.