Implementation of Convolutional Neural Network CNN Algorithm to Detect Coffe Fruit Maturity

Abstract

Fruit ripeness detection is important in the agriculture and food processing industries to ensure optimal product quality. Proper fruit ripeness can affect flavour, texture and nutrition, making it a key focus in production process monitoring and control. The fruit ripeness detection process still needs to be done manually, which can be inefficient and inaccurate. This research aims to address these challenges by implementing the CNN algorithm with VGG-19 architecture to detect coffee fruit ripeness automatically. The process involves collecting datasets of fruit images with various ripeness levels, image pre-processing including cropping and resizing, training the CNN VGG-19 model with feature learning and hyperparameter optimisation and evaluating model performance using a confusion matrix. This experiment aims to evaluate the model's performance in detecting fruit ripeness and measure the speed and efficiency of the CNN-based detection system with VGG-19 architecture. The results of this research are expected to help develop a better system for identifying fruit ripeness.