Enhancing Mango Quality Evaluation: Utilizing an MLP Model for Five-Class Severity Grading

The examination of mango quality is by all means used in the agribusiness chain that links the growers to the consumers, and both benefit from that. The traditional quality assessment looking much more manual and subjective rather than consistent and efficient is a real challenge among the others. The study proposes a breakthrough automatic technique for grading mangoes based on a multi-layer perceptron (MLP) synthesis, which effortlessly categorizes mangoes within five proxies of quality, with high reliability and accuracy. The dataset of mangoes we are harnessing, which is all with quality grades annotated, has through machine learning we trained an MLP model, and its architecture was optimized so that it can effectively identify and classify various shades in the quality of mangoes. The model exhibited a very high accuracy rate of $95.7 \%$ across grades on quality consistent with other current models and methods. The models submitted are proof of grading fruit for quality as MLP models have the potential of automation and precision, offering substantial benefits such as consistency, objectivity, and scalability. More than just helping the progress of agricultural technology, this research makes it possible for future quality control evaluation based on machine learning for every horticultural-related product. Findings confirm the usefulness of the model for cultivators, exporters, and retailers in increasing the management of quality control mechanisms, which consequently helps get rid of waste and improve the client’s satisfaction.