Penelitian untuk Aplikasi Mobile Munsell Soil Color Chart Berbasis Android Menggunakan Histogram Ruang Citra HVC dengan Klasifikasi k-NN

Abstract

Soil color determination is a challenging task because of the various types of soil and the high similarity of soil color. Munsell Soil Color Chart (MSCC) is used to overcome this issue. The purpose of this research is to develop an androidbased mobile application to identify the major color of soil according to MSCC. The application utilized HVC color model with hue, value, and chroma color components. The hue color component had been the main concern of this research, although the value and chorma color components were also observed. The training data were 259 images from MSCC, with 1600 pixels each. Hue color data from these pixels were analyzed in the form of histogram and classified with k-NN. The results showed that the highest classification accuracy of hue value was 45% at k=5, this accuracy may be improved by utilizing the soil color data that are available online.