Development of an Abaca Fiber Automated Grading System Using Computer Vision and Deep Convolutional Neural Network

This study introduces a device that can be used in classifying the eight normal hand-stripped of the abaca fiber applying the concept of Convolutional Neural Network (CNN)-MobileNetV2 algorithm. In this study., 100 sample images of abaca fiber for each class were used in training for a total of 800 sample images. The gathered data were split into a training and validation dataset using an 80/20 ratio in which 80% of the collected samples were used to train the network and 20% of it was used to validate if the trained model is within acceptable values. Using the Raspberry Pi 4, the device was made possible. Raspberry Pi 4 is a microcomputer that supports functional applications through the integration of hardware and software components. Upon testing., the developed system obtained an overall accuracy of 85%. The implementation of this device would be of great help to abaca fiber traders, producers., and Grading and Bailing Establishments.