Sistem Quality Check Box Power Emergency Menggunakan Image Processing Dengan Metode Gray Level Co–Occurrence Matrix (GLCM).

Firmansyah, Fariz Akbar (2022) Sistem Quality Check Box Power Emergency Menggunakan Image Processing Dengan Metode Gray Level Co–Occurrence Matrix (GLCM). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 10311810000037-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

PT. MBA memproduksi produk power emergency di Beji Pasuruan. Dalam proses produksi power emergency dilakukan sortir pengawasan produk akhir secara manual dengan manusia sebagai tenaga kerjanya. Dalam sehari ditemukan produk cacat dari 50-75 produksi per hari terdapat rata-rata sebanyak 3 sampai 5 produk cacat yang lolos sortir ke area good quality. Untuk meminimalisir terjadinya produk lolos sortir secara terus menerus, maka diperlukan suatu sistem yang bisa menggantikan tenaga manusia sebagai pengganti proses sortir, maka dari itu dirancanglah sebuah Sistem Quality Check Product secara otomatis menggunakan image processing dengan (GLCM) Metode Gray Level Co-Occurrence Matrix. GLCM merupakan Metode ekstraksi fitur tekstur dari object dengan cara men-ekstraksi fitur homogeneity, energy, contrast, dissimilarity, dan ASM (Angular Second Moment). Lalu hasil dari ekstraksi fitur tersebut diklasifikasi dengan metode Support Vector Machine (SVM) untuk memisahkan object box kemasan power emergency ke dalam 2 kelas yaitu kelas Bad Box dan Good Box. Apabila sistem mendeteksi adanya indikasi cacat, rusak, berlubang, dan sobek maka sistem akan melakukan proses reject dan produk di klasifikasikan pada kelas Bad Box. Pada Proyek Akhir ini menggunakan 400 dataset Box dengan masing-masing 200 dataset Good Box dan 200 dataset Bad Box. Metode klasifikasi SVM ( Support Vector Machine ) menggunakan kernel Linear dengan parameter terbaik berada pada nilai C = 1 dan Gamma = 0,01 dan menghasilkan akurasi model SVM kernel Linear sebesar 0,9625, Hasil dari pengujian sistem quality check dengan metode GLCM sebagai ekstraksi fitur, menghasilkan akurasi sistem pada pengujian dengan 40 object Random Box menghasilkan tingkat akurasi sebesar 90%, presisi sebesar 86% dan recall sebesar 95%.
====================================================================================================================================
PT. MBA produces a emergency power in Beji Pasuruan. In the production process of emergency power, manual sorting of final product supervision is carried out with humans as the workforce. In a day found defective products from 50-75 production per day there are an average of 3 to 5 defective products that pass the sorting to the good quality area. To minimize the occurrence of continuously passing products, we need a system that can replace human labor as a substitute for the sorting process, therefore a Quality Check Product System is designed automatically using image processing (GLCM) Gray Level Co-Occurrence Matrix method. GLCM is a method of extracting texture features from objects by extracting features of homogeneity, energy, contrast, dissimilarity, and ASM (Angular Second Moment). Then the results of the feature extraction are classified using the Support Vector Machine (SVM) method to separate the emergency power packaging object box into 2 classes, namely the Bad Box and Good Box classes. If the system detects an indication of defects, damage, holes, and tears, the system will reject the product and the product will be classified in the Bad Box class. In this final project, 400 Box datasets are used with 200 Good Box datasets and 200 Bad Box datasets respectively. The SVM (Support Vector Machine) classification method uses a Linear kernel with the best parameters being C = 1 and Gamma = 0.01 and produces an accuracy of the SVM kernel Linear model of 0.9625. The results of the quality check system test using the GLCM method as feature extraction , resulting in system accuracy in testing with 40 Random Box objects resulting in an accuracy rate of 90%, precision of 86% and recall of 95%.

Item Type:	Thesis (Other)
Additional Information:	RSEO 658.562 Fir s-1 2022
Uncontrolled Keywords:	Sortir, Quality Check, Image Processing, GLCM (Gray Level Co-Occurrence Matrix), SVM (Support Vector Machine). Sorted, Quality Check, Image Processing
Subjects:	T Technology > TS Manufactures > TS156 Quality Control. QFD. Taguchi methods (Quality control)
Divisions:	Faculty of Vocational > 36304-Automation Electronic Engineering
Depositing User:	Mr. Marsudiyana -
Date Deposited:	23 Apr 2026 08:05
Last Modified:	23 Apr 2026 08:05
URI:	http://repository.its.ac.id/id/eprint/132896

Actions (login required)

View Item