|
|
|
Mapping From Space
|
Present situation of spaceborne SAR
Single parameter SAR system
Single parameter SAR produces monochromatic image. In order to draw a maximum information from the single parameter SAR data, following methods are occasionally used under the favorable condition.
- Texture analysis sometimes are used for single parameter image. On the contrary, optical sensor such as MSS and MESSR have been working for object indentification and ensure a certain extent of the good results so far using three or more spectrum Channels. According to our experiment on texture analysis of SAR image by using co-occurrence Matrices, forest, vegetation, and cultivated area can be classified with the insufficient accuracy compared to the optical data. (1)
- Multi-temporal data can be extracted form single parameter SAR data of different season or different crop period. SAR data is adequate for this purpose because of the all weather capability. We can observe SAR data at exact time of period without any delay caused by
Weather conditions. It is useful for identifying the forest and crop by multi-temporal data.
- Direction of illumination is fixed to the descending and ascending spacecraft orbit. Theoretically it is possible to use these two direction as long as the satellite power is enough to make SAR observation at night orbit. SAR image is very sensitive to the incident angle to the target, so that in the mountainous area the image of different direction of illumination contains different information and shows quite dissimilar image.
- Incident angle can be varied in orbit by some SAR system. In case of resent scheduled satellites of E-ERS-1 and J-ERS-1 are fixed off-nadir angle system, however, their off-nadir angle are 20.5 degree and 35 degree respectively. It may be possible to produce sterographic image from these two SAR data. (2)
Particular characteristics of SAR data and its drawbacks.
- Rain effect on SAR observation.
All - weather capabilities of SAR is one of the most distinguished and favorite characteristic. Meteoroligical radar of wavelength of S-band or shorter observes rainfall intensity as a routine weather observation. In Japan C-band weather radar is most common. It is natural to be influenced by heavy rainfall for C-and X-band SAR. According to simple calculation of back scattering echo intensity from heavy rain up to 4-mm/hr, rain echo at L-and P-band SAR is very weak and no rain image will appear. We may safely say that the rainfall rate of 400 mm/hr is not unusual in the trop9ics. (3) At C-band, torrential rain over 100 mm/hr will mask the surface image. At x-band the intensity of rain echo is more strong than C-band and that of 20 to 30 mm/hr or more will be comparable intensity to that of natural ground echoes.
- Attenuation
Radar wave attenuation caused by meteorological particles is negligible for cloud, fog and drizzle as to the frequency in longer than X-band. In case of radar wave is propagating about 10 km in slant range through the torrential rain of 25 to 400 mm/hr, X-band will be suffered by 5 to 70 dB of attenuation. C-band about 1 to
12 dB, and L-band and P-band will be negligible less than 1 dB.
- Speckle noise is an inherent nature of active microwave system and is one big drawbacks for SAR images. Usually speckle noise is reduced at the sacrifice of spatial resolution.
- Side looking system measures the slant range to the target rather than look angle. Image distortion due to side looking system is very important nature when incident angle is small. They are foreshortening, shadow and layover.
- The large power consumption, large antenna size and heavy weight are required for multi-channel SAR system because the transmitter, receiver and antenna must be equipped separately for each frequencies. In some antenna design a common antenna reflector is used.
Future trend of spaceborne SAR (4)
From the reasons given in the preceding section, polarimetric SAR is considered one of the ideal SAR system, which include multifrequency and multi-polarization. The number of channels are decided by the combinhation of these parameters. We can get color image like the optical sensor, however, information contained in polarimetric SAR the optical sensor , however, information contained in polarimetric SAR data is different from that of optical image, development of new algorithm to analyze the data will be required. Before we reach to
this ideal polarimetric SAR system, there will be middle step of multi-channel SAR, for instance, two frequency of L-and C-band, dual polarization of HH and VV will have 4 channel information. The variable off-nadir angle system contains more information about terrains height.
Conclusion
Among many favorite characteristics such as all-weather capabilities, there are some disadvantages in connection with SAR observations which is apt to be hidden in brilliant characteristics of which is missing at optical sensors. The substantial difference between SAR and optical remote sensing system comes from following three characteristic features of the SAR:1) Active sensor, 2) Microwave sensor, 3) Side looking system. Table 2 gives the summaries of SAR features.
Table 2. Merit and demerit of SAR system
| System |
Merit |
Demerit |
| Active sensor |
Independence of sun angle |
Large power consumption heavy weight Speckle noise |
| Microwave sensor |
All - weather capability Penetration into target Sensitive to soil
moisture |
Large anatenna size |
| Side looking system |
Sensitive to surface roughness and inclination |
Distortion of layover, foreshortening, shadow |
References
- Kimura, H., T. Iijima and N. Kodaira: 1984 "Experiment on texture analysis of SAR image by using co-occurence matrices."Proc. the 4th Japanese Conference on Remote Sensing. Remote Sensing Society of Japan.
- Kimura, H.: 1990 "Geometry of Stereo SAR" Introduction of `Remote Sensing, Asakurashoten, P. 301-304.
- Moriya, T. et al.: 1989 "Rainfall characteristic in tropics and the observation method" URSI Commission F, No. 339.
- Evans, D. L. et al.: 1988 "Radar Polarimetry: anaalysis tools and applications" IEEE Transactions on Geoscience and Remote Sensing, 26-6. p. 774-789.
|
|
|
|
|
|
|
