Penerapan Neural Network untuk Klasifkasi Kerusakan Mutu Tomat

Abstract

The decrease in quality and productivity of tomatoes is caused by high rainfall, bad weather and cultivation so that the tomatoes become rotten, cracked, and spotting occurs. The government is trying to provide training to improve the quality of tomatoes for farmers. However, the training was not effective so the researchers helped create a system that was able to educate farmers in the classification of damage to tomato quality. This system serves to facilitate farmers in recognizing tomato damage thereby reducing the risk of crop failure. In this study, the classification method used is backpropagation with 7 input parameters. The input consists of morphological and texture features. The output of this classification system consists of 3 classes are blossom end rot, fruit cracking and fruit spots caused by bacterial specks. The best accuracy level of the system in classifying tomato quality damage in the training process is 89.04% and testing is 81.11%.