Image Classification of the Fertility Level of Chili Using Convolutional Neural Network

Potential exports from Indonesia's economy include agricultural products. Red chilies, especially the curly red chili, are one possible horticultural product that might be produced. For the longest time, the process of sorting chili by various firms in the processing sector, chili exporters, and chili farmers has traditionally been done manually and includes several human resources as the decision-maker. The process of sorting chili manually has many disadvantages, such as the long time required, human visual limitations, and the psychic conditions of the observer. The development of machine learning image processing technology makes it possible to sort agricultural products and plantations automatically. One of the deep machine learning methods techniques is VGG16 models from the Convolutional Neural Network method which currently has the most significant result in image recognition is VGG16 models. The classification accuracy value obtained is 93.3% with kernel size 3x3 and batch size 32 at epochs 50 using the RGB input image.