Search

This paper reviews the model-based scattering power decomposition methods of fully polarimetric SAR data. Starting from Freeman & Durden 3-component decomposition, comparison of 4-component decompositions are given in terms of available information included in the coherency matrix. The final general 4-component decomposition method with unitary transformation (G4U) employs 100% polarimetric information. [ABSTRACT FROM PUBLISHER]

As one of the image processing method of radar sensor, it is most typical method which employs synthetic aperture radar (SAR) processing. In this paper we investigate radar image of scattering models by using an extended idea of array-factor (AF) in antenna theory. Applying phase difference between each element of the array becomes so-called phased array beam-forming and weighting to each element becomes adaptive beam processing [1], [2]. Also, regarding AF as near sources, it has been directly applied to far-field transforming algorithm [3]. Since AF is considered as a sum of delayed phase signals from each antenna element, we can construct image of the signal from the sum which depends on degree of phase matching between the receiving signal and its coordinate. This procedure is a kind of aperture synthesis, also known as delay and sum (DAS), especially in the field of ultrasound medical imaging and acoustics [4]. [ABSTRACT FROM PUBLISHER]

This paper proposes a simple algorithm to precisely detect man-made objects with any alignment to radar illumination direction. Here, a couple of polarimetric indices, i.e. the copol. correlation coefficient in LR circular polarization basis, and its normalized one are utilized in the algorithm, since the former is emphasized in reflection symmetry (RS) as normally aligned man-made objects area, whereas the latter is enhanced in Non-RS as obliquely aligned man-made objects area. The appropriate thresholds for switching the polarimetric indices are determined by considering the result of polarimetric scattering analysis for simplified man-made objects model. The Finite-Difference Time-Domain (FDTD) method is utilized for the analysis. It is finally confirmed from the result of the PolSAR image analysis that the proposed algorithm is very effective for detecting the man-made objects not only normally but also obliquely alignment to the radar illumination direction. [ABSTRACT FROM PUBLISHER]

Speckle noise makes the synthetic aperture radar (SAR) image interpretation difficult and reduces the accuracy of the rapid disaster assessment. This paper introduces the influence of speckle filtering methods on the polarimetric SAR data interpretation of damages. Four types of speckle filtering methods, such as boxcar filter, refined Lee filter, Lee sigma filter, and sequential filter, are investigated to determine the best filtering options for detection of damaged areas. As a result, Lee sigma filter exhibits the best performance in the analysis of the damaged areas. The sequential filtering approach also shows a good performance, which is comparable with that of Lee sigma filter, with minimum computational cost. [ABSTRACT FROM PUBLISHER]

Polarimetric radar has a potential ability to retrieve forest parameters such as the moisture content and the biomass. Accurate retrieval of those parameters is the key to understand forest ecosystem and to prevent global climate change. For those purposes, this research aims to establish a quantitative approach to the forest parameter retrieval using polarimetric radar. To achieve this, we need a scattering model which describes scattering from forested terrain. The scattering mechanism has been extensively investigated, and several scattering models have been proposed. Although those models are validated with a field experiment, difficulties arise because those models require many input parameters which are difficult to measure accurately over widespread area. Therefore, a relatively simple scattering model is necessary, and carefully controlled experiment is essential to validate the model. In this research, we employ the discrete scatterer model (DSM) described in [1] as the forest scattering model. Using this model, we pointed out that fully polarimetric observation has the potential ability to retrieve the moisture content of forested terrain [2]. This paper aims to prove this theoretical result through an indoor fully polarimetric measurement. [ABSTRACT FROM PUBLISHER]

In this paper, an advanced version of the hybrid Freeman/eigenvalue decomposition technique for land parameter extraction is presented, denoted as G3H. It is shown that vegetation areas and oriented objects such as urban building areas are well discriminated by the proposed technique as compared to the existing Cloude hybrid decomposition technique, denoted as F3Hc [ABSTRACT FROM PUBLISHER]