A 4,310-Image Dataset Of Healthy (1,995) And Diseased (2,315) Hazelnut Clusters (Corylus Avellana L.) For Image Processing

This dataset was created to support <strong>image processing–based analysis and classification</strong> of hazelnut clusters (<em>çotanak</em>) exhibiting healthy and early-stage deterioration characteristics in <em>Corylus avellana L.</em> The dataset consists of a total of <strong>4,310 RGB images</strong>, including <strong>1,995 Healthy</strong> and <strong>2,315 Diseased</strong> hazelnut cluster samples, collected under real field conditions.

All images focus exclusively on hazelnut clusters at different developmental stages. The Diseased class includes clusters showing early deterioration symptoms such as morphological deformation, color variation, and structural irregularities, while the Healthy class represents visually intact clusters with normal morphological characteristics. Images were captured under natural illumination, with varying viewpoints and background complexity, reflecting realistic agricultural production environments.

The dataset is suitable for classical <strong>image processing and image-based classification</strong> tasks, including color space analysis, texture analysis, morphological operations, edge detection, segmentation, and feature-based classification of hazelnut cluster deterioration. The natural variability present in the images enables robust evaluation of image processing algorithms under real-world conditions.

This dataset is provided as an <strong>open-access resource for research and educational purposes</strong>, aiming to facilitate image processing studies focused on the early detection and classification of deterioration in hazelnut clusters.