Scatterometers equipped with C-bands are deployed along the coast surface for measuring ocean and sea surface winds and roughness. Deployed in low orbital satellites, the electromagnetic signals transmitted are scattered at the ocean’s surface from which the intensity and wind direction are detected. Wind intensity is impacted by different features such as salinity, scattering index, obstacles, etc., resulting in erroneous predictions. This article introduces an Attuned Data Extraction Method (ADEM) for detecting precise wind intensity and direction. The fore-mentioned errors are addressed using multimodal data fusion to prevent the density seizure problem. This density seizure is caused due to inappropriate/ irrelevant sensing. The above classification uses a random forest learning paradigm for each sensing instance. The classification refers to the inappropriate and irrelevant data observed during speed estimation. The classification is necessary to balance the variations in wind speed and intensity observed from different points. The unclassified data is neglected from the fusion process, preventing errors in the forecast. Besides, the fusion is performed in two distinct levels: extracted and attuned. In the extracted fusion data, the classified data is exploited without alignment; the attuned fusion requires error correction to improve the precision. The joint fusion data scales are utilized for improving the sensing device data consistency with less computing time and errors.
本站僅提供期刊文獻檢索。
