Polarimetric observation by PALSAR

PALSAR is the first satellite-borne L-band synthetic aperture radar capable of fully polarimetric observation. Polarization is an index describing the characteristics of electromagnetic waves and is used to express the orientation of the electric field. When the electric field is vertical, the electromagnetic wave is said to be vertically polarized. When the electric field is horizontal, the electromagnetic wave is said to be horizontally polarized. Elliptically and circularly polarized waves are formed by combining the vertically and horizontally polarized waves. Figure 1 (i) plots the locus of an elliptically polarized wave projected on a transverse plane at a fixed location along the direction of propagation. Typical polarizations are horizontal/vertical polarization, linear 45/-45° polarization, and left/right circular polarization (Fig. 1 (ii)). A synthetic aperture radar is usually designed to transmit and receive either horizontally or vertically polarized signals. For example, the Japan Earth Resources Satellite (JERS) 1 SAR utilized horizontally polarized signals. However, PALSAR can transmit and receive both horizontally or vertically polarized signals. Thus, a combined polarization radar image can be HH (horizontal transmitting, horizontal receiving), VV (vertical transmitting, vertical receiving), HV (horizontal transmitting, vertical receiving) or VH (the reverse of HV). Since the scattering depends on the polarization properties of the target (Fig. 1 (iii)), the different scattering patterns among polarizations can be observed from a PALSAR polarimetric image.

Fig. 1 Polarization of electromagnetic signal

Fig. 1 Polarization of electromagnetic signal
(Click to View Enlarged Image)


Figure 2 presents HH, HV and VH polarimetric images and RGB color composite images of Tomakomai, Japan, acquired by PALSAR using H/V polarization on August 19, 2006. This area consists of cultivated land (upper area), forest areas (middle area) and the Pacific Ocean (lower area). Although it is difficult to recognize the difference among HH, HV, and VV from the three single polarimetric images on the left, the RGB color composite image represents the scattering properties of each area as color differences.

Figure 2 presents HH, HV and VH polarimetric images and RGB color composite images of Tomakomai, Japan, acquired by PALSAR using H/V polarization on August 19, 2006. Figure 2 presents HH polarimetric images of Tomakomai, Japan, acquired by PALSAR using H/V polarization on August 19, 2006. Figure 2 presents HV polarimetric images of Tomakomai, Japan, acquired by PALSAR using H/V polarization on August 19, 2006. Figure 2 presents VV polarimetric images of Tomakomai, Japan, acquired by PALSAR using H/V polarization on August 19, 2006. Figure 2 presents RGB color composite images of Tomakomai, Japan, acquired by PALSAR using H/V polarization on August 19, 2006.

©METI, JAXA

Fig. 2. PALSAR polarimetric image
(Click to View Enlarged Image)


Figure 3 presents RGB color composite images acquired with 45/-45 linear polarization and left-/right-hand circular polarization. PALSAR cannot perform polarimetric observation using 45/-45 linear polarization and left-/right-hand circular polarization. Thus, these images were numerically constructed using observed H/V polarization data. Since the variation of color can be confirmed by comparing the three color composite images, the selection of polarization is an important factor when analyzing a targets polarimetric scattering features.

Fig. 3. Color images composed using 45/-45 linear polarization and left-/right-hand circular polarization Fig. 3. Color images composed using 45/-45 linear polarization Fig. 3. Color images composed using left-/right-hand circular polarization

©METI, JAXA

Fig. 3. Color images composed using 45/-45 linear polarization and left-/right-hand circular polarization.
(Click to View Enlarged Image)


Figure 4 depicts the cultivated area, residential area, and forest area extracted from the RGB color composite images in Figs. 2 and 3. The green region, where the cross-polarized component occurs, shifts with the change of polarization. Cross-polarized component generation is related to volume scattering from H/V polarized signals, double-bounce scattering from 45/-45 linearly polarized signals, and surface scattering from left-/right-hand circularly polarized signals. This information is used for terrain and land-use classification.

Fig. 4. Polarimetric images of cultivated, residential and forest areas. Fig. 4. Polarimetric images of cultivated areas Fully polarimetric data (H/V polarized signals). Fig. 4. Polarimetric images of cultivated areas Fully polarimetric data (45/-45 linearly polarized signals). Fig. 4. Polarimetric images of cultivated areas Fully polarimetric data (left-/right-hand circularly polarized signals). Fig. 4. Polarimetric images of residential areas Fully polarimetric data (H/V polarized signals). Fig. 4. Polarimetric images of residential areas Fully polarimetric data (45/-45 linearly polarized signals). Fig. 4. Polarimetric images of residential areas Fully polarimetric data (left-/right-hand circularly polarized signals). Fig. 4. Polarimetric images of forest areas Fully polarimetric data (H/V polarized signals). Fig. 4. Polarimetric images of forest areas Fully polarimetric data (45/-45 linearly polarized signals). Fig. 4. Polarimetric images of forest areas Fully polarimetric data (left-/right-hand circularly polarized signals).

©METI, JAXA

Fig. 4. Polarimetric images of cultivated, residential and forest areas.
(Click to View Enlarged Image)

Fully polarimetric data contains more useful information than single polarimetric data. This advantage is expected be utilized in various remote sensing applications.

©JAXA EORC

To top

JAXA,EORC Earth Observation Research Center
Copyright©1997- JAXA EORC All rights reserved.