Climacteric Defective Sapota Identification using Machine Learning-A Novel method in Segregating Latex Oozing and Rot using Augmented Images

Climacteric fruits, such as Sapota, are susceptible to post-harvest losses if not managed optimally. Manual inspection for fruit quality is time-consuming and prone to errors. To address this, we propose an image-based approach for classifying Sapota fruits into good, latex-oozed, and rot categories. The methodology involves sequential image filtering (Median, Bilateral, and Gabor) for enhancement, followed by marker-based morphological updated watershed segmentation to isolate fruit regions. Non-Parametric Naïve Bayes (NPNB) classification achieves a high accuracy of 98% in distinguishing fruit conditions, outperforming Support Vector Machine (SVM) and K-Nearest Neighbors (KNN) methods. This research offers a potential solution for improving fruit quality assessment efficiency and reducing post-harvest losses.