SPADE OC2 Mediterranean Watermelon crop dataset (W2) - VITAFields – Visual Intelligence for Targeted Agriculture

This comprehensive dataset contains high-resolution aerial and drone imagery with corresponding annotations for agricultural analysis, specifically focusing on watermelon at fully grown stage. Provided as a single, complete release, it encompasses raw image captures, camera parameters, and direct annotations, alongside fully processed RGB and 4-band multispectral orthomosaics and annotated orthomosaic image tiles. This unified collection of both raw and processed data is ideal for training computer vision models directly on map-based imagery for field-scale semantic segmentation, weed and crop classification, and precise spatial analysis. Dataset Characteristics Target Crop: WatermelonGrowth Stage: Fully Grown / Mature (Full Leaves / Complete Canopy)Image Types: RGB and Multispectral (4-Band) Technical Specifications 179 Total RGB Images 5280 x 3956 pixels 716 Total Multispectral Images 2592 x 1944 pixels (Green, Red, Red Edge, and Near-Infrared) 4474 Total Orthomosaic Tiles 512 x 512 pixels (RGB and 4-Band Multispectral) 1 RGB Orthomosaic 28650 x 31715 pixels 4-Band Multispectral Orthomosaic 28650 x 31715 pixels (Green, Red, Red Edge, and Near-Infrared) Annotations JSON FIles in LabelMe Format (Polygons) Masks Binary images PNG format Camera Parameters JSON Files in ODM format