Improved Detection of Macroalgal Bloom from Sentinel-1 Images Through Ship Masking

Abstract

Macroalgal bloom (MAB) is a global issue which affects the life and economy of the people in coastal regions. In 2008 world largest MAB occurred in the Yellow Sea nearby Qingdao occupying very large area, and was piled up in huge amount which incurred a lot of expense and labour for the cleaning up process. Afterwards, it is being occurred there every year with successive growth and coverage. Therefore, it became necessary to monitor MABs continuously from their generation period to dissipation for taking remedial measures in time. However, monitoring MABs by direct observation or sample collection over the wide areas of any sea is not feasible considering effort, time and money. Therefore, numerous researchworks have been done for the detection of macroalgal blooms in the seas. However, most of those were done by using different optical satellites which are frequently hindered by weather hazards like cloud, sea fog, storm, etc., especially in the Yellow Sea. Moreover, being optical in nature those satellites also lack data during night. These drawbacks can be minimized by using Synthetic Aperture Radar (SAR) which can penetrate these weather hazards, and being active type of remote sensing can be also be operated at night. However, there are not much research works for developing methods or algorithms for MAB detection from SAR images. Considering these here we have applied gray-level co-occurrence matrix (GLCM)