Identification of Banana Maturity (Musa paradisiaca) with Artificial Neural Network

Identification of banana maturity is still done manually by the farmers. This leads to a lack of accuracy of identification in large amounts. This study was conducted to build a system capable of identifying ripeness of bananas using neural networks with image processing. The input images were bananas which were identified based on color histogram (RGB) and thent the program wasbuilt using visual basic. The system was trained with some banana image data that have different values in each classification in order to identify the level of maturity of the bananas with backpropagation method. From the identification that had been done it was found that the success rate was 100% for raw banana, almost ripe was 80%, ripe was 100% and over ripe was 100%. The identification results wass influenced by the way of taking pictures of fruit. Keyword : Artificial neural network, back propagation, banana, identification, Image processing,