| dc.description.abstract | The leaf color of rice is closely associated with the adequacy of nitrogen (N) level of the ground. N deficiency symptoms that are most obvious and commonly seen is the reduction of green color of the leaf. This research developed a mobile application to identify the color of rice leaf and to determine the appropriate fertilizer. The application was built by using the histogram feature of several color components namely Red, Green, Blue in the RGB color space, Hue, Saturation Value in the HSV color space, and grayscale. K-Nearest Neighbor was choosen as the classification method. This research showed the highest average accuracy of 90.63% on the G (green), V (Value) and grayscale color component. The k-NN classification method produced the highest average accuracy of 72.62% when the value of k is equal to 3. | en |
