Vegetables are food ingredients of plant origin that have a high water content, vegetables can be consumed fresh or processed into a dish. The diversity of vegetables in the world causes many classification processes for vegetables. Such as classification based on cultivation method, edible organs, botanical classification and classification based on growing conditions. In this study, a dataset of 17 types of vegetables and 2,550 images of vegetables were used. The vegetable species classification process uses the Convolutional Neural Network (CNN) algorithm because it has good ability in classifying image objects. The trial process was carried out using five smartphones with an Android-based operating system. The process of designing this android-based application uses the python programming language with the Tensor flow module for the testing and training process of data. The final result of the accuracy test on vegetables resulted in an average accuracy of recognizing the types of vegetables by 70% with one of the results of the classification test on vegetables producing the highest accuracy rate of 86%.
Copyrights © 2022