Intelligent System for Predicting Orange Fruit Ripeness by Integrating Climatic and Imaging Data using CNN Model for High-Accuracy Classification

Accurately assessing the ripeness and quality of orange fruits is crucial for optimizing harvest timing, reducing waste, and ensuring market-ready produce. This paper presents an intelligent hybrid classification system that leverages both imaging and climatic data to predict the quality of oranges using a Convolutional Neural Network (CNN) model. A dataset is utilized that includes critical features such as fruit size, weight, sweetness (Brix), acidity (pH), softness, harvest time, color, variety, and blemishes. These attributes were standardized, and categorical variables were encoded for integration into a predictive model. The CNN architecture was specifically designed to process and learn from both structured climatic data and image-based features to enhance classification accuracy. After preprocessing and feature engineering, the model achieved notable accuracy in predicting quality on a 5-point scale, aligning closely with real-world quality standards for orange ripeness. With the integration of both imaging and climatic data, the system demonstrates a viable pathway toward automated, high-accuracy ripeness assessment in agricultural applications, achieving the 90% accuracy target. The findings underline the potential for this approach to streamline agricultural operations, minimize subjective quality assessments, and support a more data-driven approach in fruit harvesting practices.