Teruhisa, Komatsu, Hashim Mazlan, Nurdin Nurjannah, Noiraksar Thidarat, Prathep Anchana, Stankovic Milica, Tong Phuoc Hoang Son, Pham Minh Thu, Cao Van Luong, Sam Wouthyzen, Phauk Sophany, Muslim Aidy M, Yahya Nurul Nadiah, Terauchi Genki, Sagawa Tatsuyuki, and Hayashizaki Ken-ichi. 2020. “Practical Mapping Methods of Seagrass Beds by Satellite Remote Sensing and Ground Truthing.” Coastal Marine Science 43(1):1–25. doi: 10.15083/00079480.
This review introduces practical mapping methods of seagrass beds by satellite remote sensing with ground-truthing surveys. It briefly explains optics for understanding how to map the seagrass beds under the sea. Ground truth data are necessarily used in classifying bottom habitats and evaluating classification accuracies. Ground-truthing surveys are classified into direct methods such as video observation and manta tow with a camera, and indirect methods such as echosounder and sidescan sonar. Seagrass remote sensing begins with relating habitats from ground truth data to pixel values of a satellite image. Since satellite images with high spatial resolution require high precision of positions of ground truth data, ground surveyors need to use GNSSs with sub-meter precision. Image processing procedures are composed of geometric correction, conversion of digital number of an image pixel to radiance or reflectance, atmospheric and water column corrections, image classification, and validation of classification results (accuracy assessment), which are simply explained. It is recommended to use Depth invariant index of Lyzenga (1981) or Bottom Reflectance index of Sagawa et al. (2010) for compensating attenuation of light through atmosphere and water column to obtain better seagrass habitat classification results.
Fig. 6. Flowchart for understanding the order of each component for analyzing a satellite image and detecting seagrass distribution. Numbers in parentheses indicate chapters or sections