Our research focuses on using machine learning and drone technology to improve yield estimation in agriculture. We introduce the "Coffee and Cashew Nut Dataset," containing 6,086 images with annotations of coffee and cashew nut crops. We collected this data from different coffee and cashew growing sites across Uganda through geo-tagged and time stamped drone imagery, capturing details about crop type and fruit maturity.
We meticulously curated and annotated the drone image dataset, involving agricultural experts for validation. This high-quality dataset is publicly available for various machine learning experiments.
Our dataset has significant implications, offering precise, rapid, and cost-effective yield estimation solutions for farmers. It supports the development of machine learning models for crop classification, detection, and yield estimation, especially when combined with vegetation indices.
The dataset enables the creation of machine learning systems to assist farmers in refining yield estimates and sales predictions by detecting and counting unripe, ripe, and spoilt fruits. It's a valuable resource for advancing agriculture in Uganda and other African nations.