Identifikasi Jenis Benih Jamur Menggunakan SOM Kohonen

ABSTRAK

Tanaman jamur dibudidayakan untuk memenuhi permintaan konsumen yang menyukai produkproduk

makanan berbahan jamur. Tanaman jamur dibudidayakan melalui benih jamur, dengan 3 jenis

yaitu : benih jamur kuping, benih jamur shintake, dan benih jamur tiram. Cara membedakan jenis

benih jamur bisa dilakukan dengan melihat tekstur benih jamur, bahan benih jamur kuping terbuat

dari serat dan campuran katul cenderung lebih halus dan berwarna coklat, sedangkan benih jamur

shintake terbuat dari biji padi, dan benih jamur tiram cenderung kasar berwarna coklat muda.

Penelitian ini bertujuan untuk membuat sistem yang digunakan untuk mengidentifikasi jenis benih

jamur menggunakan co-occurrence matrix dan SOM Kohonen. Jumlah data pelatihan yang digunakan

masing-masing cluster berjumlah 20 data, sedangkan pengujian menggunakan 30 data. Tingkat

keberhasilan kerja adalah 100% untuk benih jamur kuping, 100% untuk benih jamur shintake, dan

60% untuk jamur tiram dengan pengenalan komulatif 86,67%.

Kata Kunci : Tekstur Citra Jenis Benih Jamur, Co-occurrence matrix, SOM Kohonen, Neural

Network, Jaringan Syaraf Tiruan

ABSTRACT

Mushrooms are commonly cultivated to fulfil the need for consumption of mushroom based dishes.

Mushrooms are cultivated using seeds, there are three types of seeds, namely cloud ear fungus;

shiitake mushroom; and oyster mushroom seeds. To identify types of mushroom seeds, we can look at

the texture of the seed; cloud ear fungus seeds are made up of fibers mixed with rice bran, they tend to

be finelytextured and are brownish in color; shintake mushroom seeds look like rice grains; whereas

oyster mushroom seeds tend to be coarser in texture, and the color is light brown. This research aims

to develop a system to be used to identify types of mushroom seeds using co-occurrence matrix and

SOM Kohonen. The number of data used for each cluster is 20, whereas the test uses 30 data. The

success level is as follows, 100% for cloud ear fungus seeds; 100% for shintake mushroom seeds; and

60% for oyster mushroom seeds, with a cumulative identification success of 86.67%.

Ke ywords : Image Texture of Mushroom Seeds, Co-occurrence Matrix, SOM Kohonen, Neural Network, Artificial Nerve

Download Full PDF



LEAVE A COMMENT