Multiscale Segmentation of Polarimetric SAR Data Using Pauli Analysis Images
Image segmentation is a crucial process that affects the output of any segment-based classification method and governs the interpretation process. There are many approaches for segmentation of polarimetric SAR data, such as region growing and split-merge
- PDF / 940,869 Bytes
- 5 Pages / 547.087 x 737.008 pts Page_size
- 72 Downloads / 220 Views
Multiscale Segmentation of Polarimetric SAR Data Using Pauli Analysis Images M. Dabboor, A. Braun, and V. Karathanassi
Abstract Image segmentation is a crucial process that affects the output of any segment-based classification method and governs the interpretation process. There are many approaches for segmentation of polarimetric SAR data, such as region growing and splitmerge to name a few, that have been proposed recently. This paper presents the development of a new segmentation approach based on the dominant scattering mechanisms that contribute to the backscattering process, using Pauli analysis images as input data. After accomplishing the segmentation based on the scattering mechanism, further segmentation is performed by the calculation and segmentation of histograms into homogeneous regions. State-of-art ALOS polarimetric SAR data are used in the study area which is located in the southern United Kingdom and includes the city of Minehead.
91.1 Introduction Different methods have been proposed to analyse polarimetric SAR data such as the (a) Pauli analysis method which produces the surface, double bounce and 45◦ tilted double bounce, Papathanassiou (1999), (b) Cloude–Pottier analysis method which produces the entropy, a-angle and anisotropy images and other derivatives, Pottier and Lee (1999), (c) Freeman–Durdan analysis method which produces the
double bounce, surface and volume images, Freeman and Durdanet al. (1998), (d) analysis to sphere, diplane and helix, Hellmann (1999), (e) analysis based on the Huynen parameters, Titin-Schnaider (1999), (f) decomposition based on different combinations of entropy and anisotropy, Pottier and Durdan (1999). In a previous study, it was indicated that information provided by Cloude–Pottier analysis and the images produced by different combinations of entropy and anisotropy were crucial for the determination of the number of the scattering mechanisms which participated in a knowledge-based classification procedure, Dabboor and Karathanassi (2005). Various other segmentation approaches of polarimetric SAR data have been proposed in the literature, see for example Beaulieu and Touzi (2004), Lee et al. (2001), Grandi et al.(2001). Especially, Lee et al. (2001) in their work have proposed a segmentation approach based on 3D histograms, calculated from the Pauli analysis images. In this paper, the Pauli analysis method is investigated through the development and evaluation of a new segmentation methodology applied to Pauli analysis images. The proposed segmentation methodology focuses on the hierarchicalization of the scattering mechanisms provided by the Pauli analysis method.
91.2 Segmentation Approach 91.2.1 Pauli Analysis Method
M. Dabboor () Department of Geomatics Engineering, Schulich School of Engineering, University of Calgary, Calgary, AB T2N 1N4, Canada e-mail: [email protected]
The proposed method is based on the analysis images calculated from the Pauli decomposition. Pauli analysis is one of the basic SAR polarimetric data analysis methods, Papatha
Data Loading...
