Automated Non-Destructive Coconut Quality Grading using MRI and X-Ray Image Segmentation

Abstract

Assessment of coconut quality is traditionally performed using destructive methods such as cutting and manual inspection, which are labor-intensive and unsuitable for automated grading. This paper presents a non-destructive coconut quality evaluation system based on integrated Magnetic Resonance Imaging (MRI) and X-ray image segmentation. The proposed framework analyzes internal characteristics including water content and pulp thickness using intensity-based and edge-based image processing techniques. MRI images are utilized to identify water-rich and soft tissue regions, while X-ray images provide complementary structural information for boundary detection. Quantitative measurements are obtained from segmented regions and validated through comparison with manual observations. A rule-based classification mechanism is then applied to categorize coconuts into tender or mature stages and further determine overall quality as good or not good using a combined relationship between water content and pulp thickness. Experimental observations demonstrate consistent estimation accuracy and reliable grading performance. The explainable and modular design of the system makes it suitable for practical non-destructive agricultural quality assessment and automated sorting applications.