Classification for the fruit maturity using Neural Network

Abstract

Since lately steadily improving agricultural sector especially in the production fruit. There were various methods to improve productivity fruit production. The classification for the maturity of fruits is not easily determined. This is especially true, for some fruits whose color have no direct correlation with to its level of maturity or ripeness. The levels of maturity can be determined by human expert, however for larger quantity inspection, this method is not practical. Therefore, accurate automatic classification for fruit maturity may be advantageous for the agriculture industry. In addition, consumers in supermarkets may also benefit from this system. This project is a classification for fruit maturity using neural networks system. For this study, banana was chosen because it is easy to identify its maturity level by just looking to its colors and ease of availability. Hence the data can be collected without destroying the fruit. Multilayer Perceptron (MLP) was used to classify the samples for four types of maturity levels; under ripe, unripe, ripe and over ripe maturity level. (MLP) training algorithm was used to train the MLP network and it was shown that the network was able to produce accurately for the classification of fruit samples weather it were under ripe, unripe, ripe and over ripe.