This dataset includes annualized raster images of stinknet (Oncosiphon pilulifer (L. f.) Källersjö) classification in south central Arizona from 2017 - 2025. The mapped area is comprised of Maricopa, Pinal, and Pima counties. The random forest model was created using 2019 imagery and was applied to imagery in other years. Training location data was created with the aid of National Agriculture Imagery Program aerial images and Google Street View imagery. Classified imagery consisted of Google Satellite embeddings images and harmonic regression terms derived from a time series of Sentinel-2 Normalized Difference Yellow Index. Two raster datasets are available. One containing a binary classification for each 10 m pixel and one containing a probability of the pixel being classified as stinknet.