Classification of Onion Species in the Philippines Using a Convolutional Neural Network (Cnn)

The Onion (Allium cepa) which is locally known in the Philippines as “sibuyas” is one of the most produced crops in the Philippines. It is estimated that the production of onions in the Philippines last 2020 was 229.50 thousand metric tons. The classification and sortation of the different onion classes are by manual sortation and classification in harvesting the onions from farms. The different features of the onions may be identified through training a Convolutional Neural Network Model with image processing. This study developed an onion classification system that classifies red onion, shallots, sweet onion, and yellow onion. A Convolutional Neural Network with the VGG19 Architecture was used to train a model to classify the different types of onions. The system hardware components are the Raspberry Pi 4 Model B, Raspberry Pi camera V2, and Waveshare 7-inch LCD Touch screen. The total accuracy of the system was determined using a confusion matrix. The system obtained a 95% accuracy in classifying the onion types.