Abstract:
Monitoring of crop status coupled with knowledge of growth stages of each crop is
a requirement for sustainable agriculture management. Monitoring the crop
condition manually in the field is labour, cost and time intensive. Remote sensing
therefore provides fast, cost effective and timely tools necessary for the effective
monitoring of the crops. The acquisition of cloud free optical images in tropical
regions remains a big challenge since the cropping season is characterized by high
cloud cover. Radar images which are independent of weather conditions were
therefore preferred in this study due to their consistent acquisitions. The objective
of this study was to analyse the performance of Sentinel-1 (S-1) C-band radar
images in monitoring the maize growth by investigating the transferability of the
maize phenological characteristics from one season to the other. This was
accomplished by comparing the S-1 backscatter values for the 2015 and 2016
cropping seasons. The acquired S-1 images were divided into three, according to
the acquisition modes: ascending IW1; descending IW1; descending IW3. 18 ADC
Olngatongo Company maize fields were selected, with the principal maize growth
and development stages being defined by the BBCH scale. From the results, the
maize phenological development stages could be identified from the images. The
backscatter values for fields having coincident planting dates for both 2015 and
2016 had higher similarities compared to fields whose planting dates were far apart.
The ascending IW1 had the best results. The results were however inconclusive
since fewer images were available for 2015 for comparison with 2016. Thus, a
comparison across the entire cropping season could not be conclusively undertaken.
From the study, it was concluded that the phenological characteristics extracted for
one cropping season can establish a baseline for the monitoring of the subsequent
cropping seasons, which can be an indicator of expected yields.