Detection of Underwater Laver Cultivation Fields by Synthetic Aperture Radar: Reappraisal

摘要

In our previous studies, techniques were described to extract underwater laver cultivation nets from synthetic aperture radar (SAR) amplitude data and by polarimetric analyses. The principal theory is based on the phenomenon that the sea surface above the underwater nets becomes smooth in comparison with the deep water without nets. Little radar backscatter is expected from such smooth surfaces which can be distinguished from the surrounding rough surfaces. Cultivation nets can thus be extracted from amplitude data, and the measurement accuracy can be improved using polarimetric entropy since the entropy is high in the net areas of which the images are contaminated with random system noise; while the entropy is low in the surrounding sea dominated by surface scattering only. The theory was supported by the ALOS-PALSAR data of the Tokyo Bay under moderate wind speeds, where the currents are expected to be not very strong. In the present paper, we report the preliminary results of further study of extracting laver cultivation fields in the west coasts of Korea under different metrological conditions using ALOS-PALSAR. It was found that under low wind speeds of around 2-3 m/s the underwater cultivation nets are visible in PALSAR images of the areas where tidal currents are weak, but they are difficult to detect in the areas of strong currents. Under very low wind speeds below 1 m/s, almost no images corresponding to the cultivation nets are visible. From field survey, the cultivation nets are placed on the splashing level supported by ropes and floating rods, and they may be responsible for damping small-scale waves and/or preventing their propagation. Thus, this preliminary result suggests that the accuracy of extracting underwater laver cultivation nets by SAR strongly depends on the wind speed and tidal currents.