Identifikasi Varietas Jagung Mutiara Berdasarkan Data Citra Digital Menggunakan Algoritma K-Nearest Neighbor

Abstract

Corn is one of the most important carbohydrate-producing food crops in Indonesia besides rice and wheat. Every year Indonesia can produce a very large amount of corn. It is proven that in 2017 Indonesia can produce 22.59 million tons of corn. This number is very meaningful for the economy in Indonesia. One of the most widely grown corns in Indonesia is pearl corn. However, it is quite difficult to distinguish each variety of corn because the shape and color tend to be the same. One way to find out a corn variety is by utilizing the image processing data for each corn. One of the classification methods used is K-NN. Each corn image is taken the value of area, parimeter, width, length, metric and eccentricity using the Matlab application to determine the shape of a corn, then 250 data are collected. The data that has been collected is divided based on a ratio of 70% training data 30% test data and applying k values, namely 3, 5 and 7. The classification process uses the cloud-based Rstudio application. The results obtained, it is known that the test at k = 3 gets the highest accuracy value compared to other k values. The accuracy obtained is 93.24% followed by recall 91.89% precision 94.44% specificity 94.59% error rate 6.76% and f-measure 0.9315.

Improvement of Handwriting Javascraft Image Quality and Segmentation With Closing Morphology and Adaptive Thresholding Methods

“…Sebelum melakukan proses pengolahan citra digital, maka perlu dilakukan proses akuisisi citra terhadap objek atau aksara jawa yang akan diidentifkasi untuk mendapatkan data citra aksara jawa. Akuisisi citra adalah tahap mengambil, mengumpulkan, dan menyiapkan data citra menggunakan alat pengambilan citra dan memprosesnya sehingga menghasilkan data yang dikehendaki [3]. Citra yang dimaksud ialah gambar/foto, sedangkan maksud dari digital yaitu pengolahan gambar/foto dilakukan secara digital dengan media komputer.…”