Soil Nutrient Content Classification for Essential Oil Plants using kNN

Abstract

: Essential oils can grow well and produce good quality of essential oils if planted in an area that has sufficient nutrient content. In this study, the classification of soil nutrient content was carried out using soil images as an alternative to soil testing in the laboratory. The nutrient content identified in this study is Nitrogen, Phosphorus, and Potassium (N, P, K). The identification process begins with the extraction of soil texture features using the Gray-Level Cooccurrence Matrix (GLCM) and continues with the classification of nutrient content using k-NN. As a comparison in the calculation, the validation process used data from nutrient testing results in the laboratory. Based on the results of tests on 693 data training and 297 data testing of soil images, test results are obtained accuracy of 90.5724% for Nitrogen, 92.9293% for Phosphorus, and 91.9192% for Potassium. These results indicate that image processing in soil images can be used as an alternative in identifying soil nutrient content.