
This dataset was collected at Itto SARL almond farm located in the Fes-Meknes region of Morocco and focuses on visual symptoms associated with diseases affecting almond trees. The dataset primarily targets two major categories of plant health issues: fungal infections and insect-related infestations. It is designed to support research in plant disease detection, classification, and precision agriculture applications, particularly within computer vision and machine learning frameworks.
The dataset consists of 1,717 RGB images of almond tree leaves and branches captured under real-world field conditions. It includes 8 distinct disease classes, representing a range of observable symptoms caused by both fungal infections and insect infestations, as well as their progression stages.
Images were collected in situ using multiple mobile devices, resulting in heterogeneous image characteristics in terms of resolution, lighting conditions, focus, and background variability. This diversity reflects realistic agricultural monitoring scenarios and enhances the ecological validity of the dataset for deployment-oriented machine learning systems.
No controlled laboratory setup was used; instead, all samples were acquired under natural environmental conditions, including variations in illumination, occlusion, and leaf orientation.
The dataset is hierarchically structured according to disease etiology. Two primary categories are defined:
Within these categories, the dataset is further divided into specific disease classes, each corresponding to distinct visual symptomatology: