Abstract:
Wheat is one of the most important staple crops consumed by more than four billion people in the world.
However, its production is challenged by the impact of climate change which accounts for a 5.5 % reduction in
wheat yield and it is predicted to dwindle further by about 30 % in 2050, due to trends in temperature, precipitation, and carbon dioxide. An effective annual crop estimate is necessary not only to inform governments the
status of national food security, but also to determine the benchmark on which agricultural commodities are
priced in the market. Thus, annual crop monitoring and yield estimate is paramount to determine the amount of
wheat imports required to make up for the shortfalls in the national wheat production in South Africa, which has
been a net importer of wheat since 1998. This study aimed at investigating the most distinguishable crop
phenology for accurate winter wheat classification during the growing season from August – December 2020
using Sentinel-2 imageries and Random Forest algorithm. The winter wheat crop was more accurately identified
during the crop ‘heading’ stage in October yielding the highest user’s (75.56 %) and producer’s (92.52 %) accuracies, despite the relatively lower overall accuracy (78.14 %) compared to that of December with overall
accuracy of 83.58 % obtained during the maturity stage. This study, therefore, found that the extraction of NDVI
values of the winter wheat crop over the period of the growing season using the Sentinel-2 NDVI series method
and grouping these values into distinct classes using the K-means unsupervised clustering techniques assist to
identify the different crop phenologies based on which the winter wheat crop could be detected and mapped
accurately. The phenology-based classification of the winter wheat crop during the heading stage, reduce the
ambiguity of spectral confusion created with surrounding grass and maize crops.
