PERTANIKA JOURNAL OF SCIENCE AND TECHNOLOGY

The use of coconut in the food industry is determined by the condition of its fruit ripeness level, which is very difficult to be conducted. The suitable non-invasive sensing method is the application of computer vision. The purpose of this study is to identify the coconut ripeness level based on several parameters of fruit volume, coconut flesh thickness, and coconut flesh weight by using Artificial Neural Network (ANN) modeling. The best ANN model resulted in 14 inputs consisting of color, texture, shape and size parameters. Color features include: Red, Green, Blue, Hue, Saturation, and Intensity. Textural features include: contrast, correlation, energy, homogeneity. The shape and size parameters include area, perimeter, eccentricity, and metric. The best ANN structure consisted of 14 inputs, one hidden layer with 100 nodes, and 3 outputs of coconut ripeness (indicated by coconut water volume, coconut flesh thickness and wet weight of coconut flesh). The best ANN model produced the smallest Mean Squared Error (MSE) training and MSE testing values of 0.002155 respectively of 1 and 0.90331. Thus, computer vision and ANN models can be utilized to predict the coconut ripeness level.