A two-step approach for detecting Striga in a complex agroecological system using Sentinel-2 data

Information on weed occurrence within croplands is vital but often unavailable to support weeding practices and improve cropland productivity assessments. To date, few studies have been conducted estimate map abundances agroecological systems from spaceborne images over wide-area landscapes, particularly for the genus Striga. Therefore, this study attempts increase detection capacity of Striga at subpixel size using high-resolution imagery. In study, a two-step classification approach was used detect (Striga hermonthica) in Rongo, Kenya. Firstly, multidate multiyear Sentinel-2 (S2) data (2017 2018) were utilized non-cropland areas random forest algorithm Google Earth Engine. The class thereafter masked out single date S2 image 13th December 2017. remaining area then multiple endmember spectral mixture analysis (MESMA) infestation endmembers (EMs) obtained in-situ hyperspectral data. gathered resampled waveband configurations three representative EMs inferred, namely: (1) Striga, (2) crop other weeds, (3) soil. Overall accuracies 88% 78% pixel-based mapping achieved, respectively. Furthermore, an F-score (0.84) root mean square error (0.0075) showed that MESMA provides plausible results predicting relative abundance each pixel landscape scale. capability together with hierarchical thus proven be suited heterogenous system. These can guide adaptation, mitigation, remediation already infested areas, thereby avoiding further new croplands.