<![CDATA[The advancement of technology has driven the agricultural industry to become more advanced and modern. Distinguishing honey with nearly identical colors is a challenging task. However, The ability to differentiate the color of acacia and forest honey is the simplest approach to ensuring the authenticity and quality of honey products. This study aims to develop a honey color classification model using Multilayer Perceptron (MLP). Image data were collected from various angles under natural lighting, followed by ninety experiments using parameter combinations, including data imbalance handling methods, dense layer structures, and training settings. The results showed that the MLP model with an optimal configuration, utilizing the Adaptive Synthetic Sampling (ADASYN) method for data imbalance, achieved a validation accuracy of 90.63%. This accuracy highlights the potential of the model to support industrial automation processes in reliably distinguishing honey colors.]]>
Copyrights © 2025
