Color-Based Spot Detection Using Automatic Leaf Segmentation in Potato Plants

Abstract

Potato (Solanum tuberosum L.) is one of the world’s major food crops, playing a vital role in supporting food security and nutritional resilience. However, its productivity is often threatened by foliar diseases such as early blight and late blight, which can cause significant yield losses. This study aims to develop a lightweight and explainable classification method for detecting potato leaf diseases based on automatic leaf segmentation and color-based spot analysis. Early and accurate disease detection is essential to support preventive actions in plant protection. The proposed method integrates automatic leaf segmentation using HSV-based thresholding to isolate the leaf region from the background, followed by color-based spot detection to identify disease symptoms. Extracted features include spot area, number of detected spots, and average hue values, which were then classified into three categories (healthy, early blight, and late blight) using a rule-based approach. Validation was conducted by manually comparing classification outputs with ground truth derived from file names. The results show that the method can successfully segment potato leaves, detect spot regions, and classify disease types consistently with manual validation. Although not evaluated through large-scale statistical metrics, the findings indicate that this color-based approach provides a reliable foundation for lightweight potato leaf disease detection without requiring deep learning models.