Klasifikasi grade daun padi sebagai penentu pemupukan urea dengan metode ekstraksi fitur

Abstract

Most of Kalimantan lands are dominated by wetland such as rivers, swamps, and coasts. This wetland indirectly plays a role on hydrology processes in Kalimantan island. Kalimantan Selatan has 37.531 km2 wide area of land with 382.272 ha wide area of wetland.  One of wetland problems, generally, can be solved by planting wetland crops that can absorb nutrient on wetland. Rice is one of wetland crops which is suitable to be planted on wetland. Fertilizer requirement for rice to grow properly must be in accordance to the needs of the rice itself. Rice farmers usually use Leaf color chart tool (BWD) to estimate the fertilizer requirement for rice based on its leaf color. The result of this chart tool depends on human visual which is subjective. Therefore, this chart tool needs to be digitized and implements intelligent system so it can resolve problems caused by human vision. BWD can be build to work automatically by using Fuzzy Logic method. The determining parameters of rice leaf characteristics can be obtained from rice leaf two-dimension (2D) images by extracting the texture features based on Hue Saturation Intensity from the images. This research explains on how to determine the urea fertilizer requirement of rice based on its leaf images.