SmartBerry: Strawberry Field Image Dataset from Nigeria

The SmartBerry dataset is an original collection of strawberry images acquired under real agricultural production conditions in Nigeria as part of the SmartBerry project. The dataset contains nine classes representing healthy, diseased and stress-related strawberry leaf and fruit conditions. Unlike controlled laboratory datasets, the images capture natural variation in illumination, background, camera viewpoint, plant position, symptom appearance, image quality and partial occlusion. The dataset was developed to support research in strawberry health recognition, computer vision, machine learning and smart agriculture, particularly the evaluation of models under real-world field conditions. A subset comprising healthy leaves and leaf dark-spot symptoms has been used for controlled-to-field evaluation against the PlantVillage dataset and for label-efficient domain-adaptation experiments. The dataset accompanies the study “Cross-Dataset Strawberry Disease Recognition from Controlled to African Field Conditions.”