Budiman, Saiful Nur (2016) Sistem Pengukuran Mutu Buah Mangga Berdasarkan Kematangan, Ukuran Dan Area Bercak Menggunakan Fuzzy Inference System. Masters thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
5113201030-master-theses.pdf - Published Version Download (2MB) | Preview |
Preview |
Text
5113201030-paperpdf.pdf - Published Version Download (643kB) | Preview |
Preview |
Text
5113201030-presentationpdf.pdf - Published Version Download (1MB) | Preview |
Abstract
Mangga merupakan buah musiman yang dipanen secara serentak ketika
musim panen tiba. Grading mangga secara konvensional dengan menggunakan
tenaga manusia memerlukan waktu lama dan hasil yang tidak maksimal. Grading
dengan komputer berbasis sensor telah pula dikembangkan tetapi keberhasilannlya
tergantung pada sensor yang digunakan. Untuk mengatasinya, grading telah
dilakukan dengan CVS (Computer Vision System) dengan input berupa citra. Proses
CVS dimulai dari preprocessing, ekstraksi fitur, pelatihan dan pengenalan.
Parameter yang digunakan dalam proses grading adalah tingkat kematangan dan
kualitas suatu mangga. Tingkat kematangan sebuah mangga ditentukan oleh
perubahan warna, sedangkan untuk kualitas ditentukan oleh ukuran dan rasio
bercak. Pada penelitian ini FIS (Fuzzy Inference system) Mamdani digunakan untuk
mendapatkan hasil grading yang lebih baik berdasarkan parameter warna, ukuran,
dan bercak suatu mangga. Conveyer belt digunakan untuk mengangkut mangga
sehingga proses grading lebih cepat. Akibatnya, citra mangga mengalami motion
blur. Citra input yang mengalami motion blur disimulasikan dengan MATLAB.
Citra mangga diambil pada kondisi diam dan bergerak untuk menentukan akurasi
grading. Hasil penelitian menunjukkan bahwa pada kondisi diam diperoleh akurasi
sebesar 77%. Sementara pada kondisi bergerak akurasinya semakin menurun.
Dapat disimpulkan bahwa nilai akurasi grading mangga semakin menurun jika
derau motion blur semakin besar.
===========================================================
Mango is a seasonal fruit which is harvested simultaneously when the
harvest season arrives. Conventionally grading of mango by manpower spent much
more time and the results were not optimal. Computerized grading based on sensor
has also been developed but the results were sensor dependent. To solve the
problem, grading was performed by CVS with images as input. The process of CVS
was started from preprocessing, feature extraction, training and recognition.
Parameters used in the grading process were ripeness and quality of a mango. The
level of mango ripeness was determined by color change, while the quality was
determined by the size and the ratio of spotting. In this research, FIS Mamdani was
used for providing better grading based on colour, size, and spotting of a mango.
Conveyer belt was used to transport mango so the grading process was faster. As a
result, the mango image experienced motion blur. The motion blur image was
simulated with MATLAB. Mango images were captured at stopping and moving
condition to determine the grading accuracy. As a result, at stopping position the
accuracy was as much as 77%. Meanwhile, the accuracy was tend to decrease at
moving condition. It can be concluded that the accuracy of the mango grading
would be decreased as the increasing of motion blur noises.
| Item Type: | Thesis (Masters) |
|---|---|
| Additional Information: | RTIf 006.42 Bud a |
| Uncontrolled Keywords: | Grading Mangga, CVS, Motion blur, Fuzzy Inference System Mamdani. |
| Subjects: | T Technology > TA Engineering (General). Civil engineering (General) > TA158.7 Computer network resources |
| Divisions: | Faculty of Information and Communication Technology > Informatics > 55101-(S2) Master Thesis |
| Depositing User: | EKO BUDI RAHARJO |
| Date Deposited: | 14 Oct 2019 01:56 |
| Last Modified: | 14 Oct 2019 01:56 |
| URI: | http://repository.its.ac.id/id/eprint/71145 |
Actions (login required)
 |
View Item |