Passe, Rohandi Dianta (2017) Identifikasi Varietas Benih Cabai Berbasis Machine Vision Menggunakan Metode Artificial Neural Network (ANN). Sarjana thesis, Universitas Brawijaya.
Abstract
Penelitian ini bertujuan untuk mengetahui hubungan antara parameter gambar atau citra terhadap varietas benih cabai menggunakan metode Artificial Neural Network (ANN) dan mengidentifikasi varietas benih cabai berbasis machine vision menggunakan ANN. Penelitian ini menggunakan 5 jenis varietas benih cabai yang dilakukan proses scanning, dimana setiap 1 jenis varietas cabai tersebut dilakukan 100 kali proses scanning dalam alat scanner printer Canon MP 237 dalam format bitmap (.bmp) dan mengekstrak data citra pada program features extraction. Input yang digunakan untuk neural network prediction (NNpred) ada 2 model, yaitu model 1 dengan input 10 parameter tekstur berdasarkan red, green, blue, gray, hue, saturation hsl, saturation hsv, light, value, lab_L, lab_a, lab_b. dan model 2 dengan input 12 parameter warna berdasarkan energy, entropy, contrast, homogenity, invers, correlation, summean, variance, cluter, maximum probability. Kedua model tersebut menggunakan data total sebanyak 500 dan validasi sebesar 20%. Dari 2 model tersebut dilakukan pencarian laju pembelajaran dan momentum pada aplikasi NNpred, yaitu Model 1 ditemukan hasil terbaik pada laju pembelajaran 0,8 dengan momentum 0,8 dan didapatkan hasil terbaik ada pada parameter warna Lab_a dengan MSE validation 0,00005 dan ARE 1,30%. Model 2 didapatkan hasil terbaik pada laju pembelajaran 0,1 dengan momentum 0,9 dan didapatkan hasil terbaik tiap varietas pada tekstur energy dengan MSE validation 0,00003 dan ARE 1,09%.
English Abstract
The objective of this research is to determine the relationship between image parameters on chilli seeds varieties using Artificial Neural Network (ANN) method and to identify of chilli seeds varieties with machine vision based using ANN. This research used 5 types of chili seed varieties which are scanning process, where every 1 type of chili varieties is done 100 times scanning process in scanner tool of Canon MP 237 printer in bitmap format (.bmp) and extract image data in feature extraction program. Input used for neural network prediction (NNpred) there are 2 models, model 1 is input 10 of texture parameters based on red, green, blue, gray, hue, saturation hsl, saturation hsv, light, value, lab_L, lab_a, lab_b. and model 2 is input 12 of color parameters based on energy, entropy, contrast, homogenity, inverse, correlation, summean, variance, cluter, maximum probability. Both models use a total of 500 data and 20% validation. From the two models, the learning rate and momentum of the NNpred application were chosen. Model 1 is found the best result at learning rate 0.8 with 0.8 momentum and got the best result on color parameter Lab_a with MSE validation 0,00005 and ARE 1, 30%. Model 2 is got the best result at learning rate 0,1 with momentum 0,9 and got best result each varieties on energy texture with MSE validation 0,00003 and ARE 1,09%.
| Item Type: | Thesis (Sarjana) |
|---|---|
| Identification Number: | SKR/FTP/2017/648/051711551 |
| Uncontrolled Keywords: | Benih Cabai, Machine Vision, Artificial Neural Network (ANN) |
| Subjects: | 600 Technology (Applied sciences) > 635 Garden crops (Horticulture) > 635.6 Edible garden fruits and seeds |
| Divisions: | Fakultas Teknologi Pertanian > Keteknikan Pertanian |
| Depositing User: | Yusuf Dwi N. |
| Date Deposited: | 23 Jan 2018 06:45 |
| Last Modified: | 26 Nov 2020 04:13 |
| URI: | http://repository.ub.ac.id/id/eprint/8358 |
Actions (login required)
 |
View Item |