Coffee and Cashew Nut Dataset

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.

本研究聚焦于借助机器学习（machine learning）与无人机技术（drone technology）提升农业产量估算的精度。我们构建并发布了咖啡与腰果数据集（Coffee and Cashew Nut Dataset），该数据集包含6086张带有咖啡与腰果作物标注信息的影像。本数据集通过地理标记（geo-tagged）且带时间戳（time stamped）的无人机影像，从乌干达境内多处咖啡与腰果种植基地采集所得，涵盖了作物类型与果实成熟度的相关细节。我们对该无人机影像数据集进行了精细化的整理与标注，并邀请农业专家参与验证工作。该高质量数据集已对外开放，可用于各类机器学习相关实验。本数据集具备重要应用价值，可为农户提供精准、高效且经济的产量估算解决方案。其可支撑作物分类、目标检测与产量估算相关机器学习模型的研发，尤其在结合植被指数（vegetation indices）时效果更佳。该数据集可用于构建机器学习系统，通过检测并统计未成熟、成熟与腐烂的果实，帮助农户优化产量估算与销售预测工作。对于乌干达及其他非洲国家的农业发展而言，本数据集是一项宝贵的资源。