Machine Learning Based Analysis of Real-Time Geographical of RS Spatio-Temporal Data

Abstract

Flood disasters can be reliably monitored using remote sensing photos with great spatiotemporal resolution. However, satellite revisit periods and extreme weather limit the use of high spatial resolution images. As a result, this research provides a method for combining Landsat and MODIS pictures to produce high spatiotemporal imagery for flood disaster monitoring. Using the spatial and temporal adaptive reflectance fusion model (STARFM), the spatial and temporal reflectance unmixing model (STRUM), and three prominent algorithms of flexible spatiotemporal data fusion (FSDAF), Landsat fusion images are created by fusing MODIS and Landsat images. Then, to extract flood information, utilize a support vector machine (SVM) to classify the fusion images. Assess the accuracy of your work. Experimental results suggest that the three spatio-temporal fusion algorithms may be used to effectively monitor floods, with FSDAF's fusion results outperforming STARFM and STRUM in both study areas. The overall flood classification accuracy of the three STARFM, STRUM, and FSDAF algorithms in the Gwydir research region is 0.89, 0.90, and 0.91, respectively, with Kappa coefficients of 0.63, 0.64, and 0.67. The flood classification accuracy of the three fusion algorithms in the New Orleans research region is 0.90, 0.89, and 0.91, with Kappa values of 0.77, 0.76, and 0.81, respectively. The spatio-temporal fusion technique can be used to successfully monitor floods, according to this study.