Towards Smart Agriculture Through Auto-Quality Checker And Grader Of Mango Fruits Using Ensemble Machine Learning Model

Abstract

A variety of  Picture processing and ML technologies currently used in the systems that classify mango quality. Traditional algorithms usually work individually and aim to identify patterns of functions in data records. However, all technologies have their own strengths and limitations. This study introduces an ensemble learning frame that integrates several ML models to improve prediction accuracy. The process begins with extracting external features of mangoes with different image processing techniques. These functions combine with sensor-derived weight values ​​to form a data record. Several ML algorithms are tested on this data record to identify the most effective ones. Next, ensemble learning strategies such as baking, boosting, and stacking are used. Experimental results show that image processing methods significantly reduce errors. Furthermore, model assessments show that the stacking ensemble combined various basic and meta-learners with accuracy, recall, F1 scores, and accuracy values ​​with 0.9855, 0.9901, 0.9876 or 0.9863. These results confirm the effectiveness and reliability of the proposed approach for mango quality classification.