Automated Sustainable Grading of Chali Arecanuts: A Machine Learning Approach to Quality Assurance

This paper demonstrates an AI-driven method for automatically evaluating arecanuts using computer vision and machine learning approaches is presented in this study. Inconsistent quality assessment, work effort, and subjectivity are the common problems faced when opted for the manual grading techniques. Dataset comprising of more than 3,700 high-resolution images with white background in four quality grading was collected. Convolutional neural networks (CNNs) are used in the suggested system to evaluate important quality attributes as size, color, texture, and structural integrity. Several architectures were evaluated, including an ensemble approach that mixed Support Vector Machine (SVM) and Random Forest (RF), InceptionV3, ResNet50, EfficientNetB3. The study highlights real-world applicability by optimizing for low-resource hardware, energy economy, and operational simplicity, compared to traditional techniques, the approach provides significant improvements in processing speed and grading uniformly. Although previous studies have indicated good lab-level accuracy, issues such as field variability and a lack of standardized grading procedures still exist. The proposed work aims to improve production and guarantee consistent quality standards across farms by developing a scalable, affordable, and reliable precision agriculture solution.