| dc.description.abstract | The most important agricultural crop in southern Africa is maize, which grows on variety of environments and serves as an essential food source for the region. Most of the maize is grown in smallholder croplands both for subsistence and commercial purposes. It is one of the two main crops that are impacted by water stress globally. Therefore, determining maize water stress is essential for the development of timely response measures to boost farming production, especially on smallholder croplands. Unmanned Aerial Vehicles (UAVs) furnished by multispectral devices propose a technique aimed at spatially comprehensive data suitable to defining maize water stress at the farm scale. Therefore, this thesis intended toward assessing the use of UAVs-acquired information to quantitatively enumerate maize water stress. This overarching objective was addressed by two specific objectives which were to 1) conduct a systematic literature review of remote sensing data use in determining maize water stress at a farmstead level and 2) assess UAVs acquired data and machine learning (ML) techniques utility in estimating maize Crop Water Stress Index (CWSI) as an indicator for crop water stress and 3) estimate maize water stress across different phenological stages using UAVs acquired data in smallholder croplands. Particularly, the reviews assessed the distribution of publications, the types of methods used, and the types of results obtained, identifying gaps, challenges, and limitations associated with the remote sensing use for maize crop water use in smallholder farms. | en_US |
