Examining the Fusion and Assimilation of Landsat 8, Sentinel 1 Satellite Data and Derived Indices in Mapping Softwood Forest Vegetation (Southern Yellow Pines)
Remote sensing classification is useful in delineating and mapping of landcover types such as softwood forest species including southern yellow pines. However, the degree of accuracy in classification varies partly due to factors such as mixed pixel, factor of scale, spectral information and classification techniques. This study explored multispectral data integration to examine classification accuracies of softwood forest vegetation. The goal of the study was to classify, map and examine the accuracy of southern yellow pines (loblolly, shortleaf and virginia pines) under three dataset integration scenarios: 1) Landsat 8 only classification; 2) fusion of Landsat 8 and Sentinel 1 classification and; 3) integration of the fused Landsat 8 and sentinel 1 with satellite derived indices classification. Machine learning remote sensing classifier (random forest) was used to classify the softwood forest species. We found that the overall mapping accuracy of southern yellow pines increased by 7% when classification was performed with fused Landsat 8 and Sentinel 1 satellite data relative to classification performed with Landsat 8 satellite data alone. The integration of satellite derived indices with satellite data did not significantly improve the overall classification accuracy of southern yellow pines. The fusion of satellite spectral information could significantly improve the delineation of softwood forest vegetation.
Remote sensing|Forestry|Geographic information science
Eze Onyekachi Amadi,
"Examining the Fusion and Assimilation of Landsat 8, Sentinel 1 Satellite Data and Derived Indices in Mapping Softwood Forest Vegetation (Southern Yellow Pines)"
ETD Collection for Tennessee State University.