<![CDATA[The agricultural sector is essential to meet the world's food needs, for example, in rice farming in Indonesia, problems that occur in rice plants are usually not only caused by fertilization but also the result of various diseases. The aim of this study to classify nitrogen fertilizer requirements and plant diseases was made based on leaf color using the Convolutional Neural Network (CNN) method which can be used to increase the accuracy of observations because it is objective. This study uses the Kaggle dataset with a total of 1600 data divided by 4 criteria. The data set was then divided into 70 percent of the training section, 15 percent of the validation section, and 15 percent of the test section, then pre-processed the rice leaf image with color image features and GLCM. The preprocessing results are processed using the CNN method to provide results for detecting plant diseases and the need for appropriate nitrogen fertilization. The calculation of plant diseases using the CNN method offers the highest accuracy of 98.33%. The highest accuracy for the nitrogen requirement problem was 81.67%, but with a very low precision value of 4.55%. Calculating plant diseases using the CNN method can provide satisfactory results on rice leaf image datasets so that it can be used as a basis for improving seed quality. Keywords: CNN; images of rice leaves; machine learning]]>
Copyrights © 2023
