Siahaan., Eben Ezer (2018) Pendugaan Mutu Fisik Biji Kedelai (Glycine Max (L.) Merrill) Dengan Menggunakan Image Analysis Dan Artificial Neural Network (Ann). Sarjana thesis, Universitas Brawijaya.
Abstract
Beberapa jenis Kedelai memiliki kesamaan karakteristik yang tidak jauh berbeda dengan yang lainnya. Kesamaan karakteristik tersebut tidak cukup hanya dibedakan secara manual menggunakan indera manusia, Mutu pada Kedelai dapat dideteksi dengan pengolahan citra, maka tujuan dari penelitian ini yaitu memprediksi mutu pada Kedelai dengan metode image analysis menggunakan jaringan syaraf tiruan dan menentukan model Jaringan Syaraf Tiruan untuk memprediksi mutu pada Kedelai. Akuisisi citra menggunakan scanner canon pixma MP237 dengan resolusi 600 dpi sebanyak 400 gambar masing-masing varietas biji kedelai sebanyak 100 gambar. Pengolahan citra dilakukan untuk memperbaiki resolusi citra sehingga gambar yang dipakai memiliki resolusi sama yaitu 420 x 405 piksel dengan format bitmap (.bmp) dan mengestrak data citra pada program feature extraction. Setelah itu didapatkan kombinasi nilai warna nilai red, green, blue, gray, hue, s(HSL), s(HSV), L*, a*, b* dari energy, entrophy, contrast, homogenity, invers, corelation, sum mean, variance, cluster, dan max probability. Selanjutnya proses penyusunan model dan NN-pred (Neural Network Prediction). Dalam penggunaan ANN dibuat dua model yaitu model 1 dengan 10 parameter input dan model 2 dengan 12 parameter input. Kedua model dipakai untuk menghasilkan nilai MSE terkecil. Hasil penelitian menunjukkan untuk model 1 laju pembelajaran 0,7 dan momentum 0,3 sedangkan model 2 laju pembelajaran 0,2 dan momentum 0,2. kedua model tersebut menggunakan data total sebanyak 400 dan menggunakan validasi 10% sehingga diperoleh hasil untuk model 1 nilai MSE validasi terkecil tiap varietas biji kedelai terdapat pada warna lab_L sebesar 0.001 dan ARE 2,28%, pada model 2 nilai MSE validasi terkecil untuk tiap varietas biji kedelai terdapat pada tekstur max.probability sebesar 0.001 dan ARE 3,05%.
English Abstract
Some types of soybeans have similar characteristics that are not much different from the others. The similarity of characteristics is not enough to only be manually differentiated using the human senses. Quality in soybeans can be detected by image processing. The purpose of this study is to predict quality in soybeans using the image analysis method using artificial neural networks and determine models of Artificial Neural Networks to predict quality at soybean. Image acquisition using a Canon Pixma MP237 scanner with a resolution of 600 dpi as many as 400 images of each variety of soybean seeds as many as 100 images. Image processing is done to improve image resolution so that the image used has the same resolution ie 420 x 405 pixels in bitmap format (.bmp) and to restore image data to the feature extraction program. After that, the combination of color values of red, green, blue, gray, hue, s (HSL), s (HSV), L *, a *, b * of energy, entrophy, contrast, homogeneity, inverse, correlation, sum mean , variance, cluster, and max probability. Next is the model compilation process and NN-pred (Neural Network Prediction). In the use of ANN two models were made, namely model 1 with 10 input parameters and model 2 with 12 input parameters. Both models are used to produce the smallest MSE value. The results of the study show for the model 1 rate 0.7 learning and 0.3 momentum while model 2 learning rate 0.2 and momentum 0.2. the two models used a total data of 400 and used 10% validation so that the results for model 1 MSE values the smallest validation of each soybean variety found in lab color of 0.001 and ARE 2.28%, in the model 2 the smallest validation MSE value for each variety soybean seeds are found in max.probability texture of 0.001 and ARE 3.05%.
| Item Type: | Thesis (Sarjana) |
|---|---|
| Identification Number: | SKR/FTP/2018/622/051900652 |
| Uncontrolled Keywords: | Artificial Neural Network, Biji Kedelai, Features Extraction, Machine Vision |
| Subjects: | 600 Technology (Applied sciences) > 664 Food technology > 664.7 Grains, other seeds, their derived products > 664.72 Milling and milling products > 664.726 Nuts, legumes, other non-cereal seeds and theri flour, meals, by-products |
| Divisions: | Fakultas Teknologi Pertanian > Keteknikan Pertanian |
| Depositing User: | Endang Susworini |
| Date Deposited: | 22 Oct 2019 01:50 |
| Last Modified: | 11 Jan 2022 03:01 |
| URI: | http://repository.ub.ac.id/id/eprint/165092 |
Preview |
Text
Eben Ezer Siahaan.pdf Download (3MB) | Preview |
Actions (login required)
 |
View Item |