Urban density estimation from polarimetric SAR images using polarization orientation angle

Abstract

In this paper, an urban density estimation algorithm from polarimetric synthetic aperture radar (SAR) images is proposed. Polarization orientation angle (POA) and four power components derived by four-component decomposition are utilized in the proposed algorithm. The dependence of the four decomposition components on POA can be lessened by rotating the elements of the coherency matrix by the POA. However, a level of POA dependence remains even after the correction. The proposed algorithm utilizes POA-corrected components, but pixels are grouped into several categories according to POA. Then, each component's scattering intensity is normalized for all pixels in each POA space. Finally, both the normalized intensity of volume scattering power (Pv) and that of the subtraction of surface scattering power (Ps) from double bounce scattering power (Pd) are combined to estimate urban density.

Global urban mapping using building density from polarimetric SAR images with POA correction

“…Fig. 3 shows the results of urban density estimation for Tokyo, New York, and Munich by using the proposed method of [1]. These estimated urban density ranges between 0 and the estimated value shown in Table II for each area.…”

“…Urban density estimation is based on the proposed method of [1]. The results after simply applying the method to PolSAR images have the same scale of urban density.…”

“…The proposed method uses T11, T22, and T33 which are the diagonal elements of the coherency matrix. According to an analysis of experimental data [1], double-bounce scattering power and volume scattering power have a positive correlation with building density, in contrast, surface scattering power has a negative correlation with building density. By considering this tendency, "T22 -T11 (+ T33)" is expected to be an effective index related to building density.…”

“…Polarimetric SAR data format consists of the complex scattering matrix given by (1) Here, for simplicity, S HV and S VH are assumed to be equivalent, and the coherency matrix is expressed by (2) The diagonal elements of the coherency matrix (T11, T22, and T33) are real numbers. T11, T22, and T33 correspond to surface scattering, double-bounce scattering, and volume scattering process, respectively.…”

“…An algorithm for estimating urban density using polarimetric synthetic aperture radar (PolSAR) data was proposed [1]. This algorithm considers the effect of building azimuth angle on PolSAR images by using Polarization Orientation Angle (POA) [2].…”