Dataset for chickpea quality classification using ESP32-CAM

Image dataset of chickpeas (Cicer arietinum) captured with an ESP32-CAM module for automated quality classification. The dataset contains 283 RGB images (800x600 px, PNG format) organized into 5 categories: Sanos/Healthy (60), Danados/Damaged (50), Humedad/Humidity (60), Piedras/Rocks (53) and Entrenamiento/Training (25, held out as a blind-test set). Images were captured between March and July 2026 under controlled lighting conditions using a fixed camera rig with a printed scale reference (13 px/mm). This dataset was used to train and validate CART (Decision Tree) classifiers described in the accompanying article and source code.