Sistem Inspeksi Visual Penempatan Label Produk Lip Cream Menggunakan Metode Deep Learning Di Pt Paragon Technology And Innovation

Triatmaja, Moch Deny (2023) Sistem Inspeksi Visual Penempatan Label Produk Lip Cream Menggunakan Metode Deep Learning Di Pt Paragon Technology And Innovation. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 10311910000073- Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 September 2025. Download (9MB) | Request a copy

Abstract

Improvement yang dilakukan PT PTI menggunakan mesin SBL (single bottom label) untuk penempelan bottom label secara in line pada line produksi VSN 03. Mesin SBL masih belum memiliki chacker hasil penempelan label. Oleh karena itu, pos operator selanjutnya perlu melakukan dua pekerjaan yaitu melakukan quality checker hasil penempelan mesin SBL dan memasukan packaging ke dusat. Proses quality checker memiliki kekurangan yaitu terjadinya human error yang disebabkan beberapa faktor diantaranya ketelitian dan ketajaman mata manusia yang berbeda-beda. Pada penelitian ini dibuat sebuah sistem mesin sortir yang dilengkapi sistem deteksi penempelan label produk lip cream menggunakan algoritma deep learning YOLO (You Only Look Once). Algoritma ini memproses gambar secara real-time pada 45 fps. YOLO mendeteksi objek dengan menggunakan unfield model dimana sebuah single convolutional network memprediksi beberapa bounding boxes (kotak pembatas) serta probabilitas kelas di dalam kotak-kotak tersebut secara bersamaan. Hasil deteksi dan klasifikasi produk akan ditampilkan pada monitor kemudian produk yang not good akan disortir dengan cara di tiup menggunakan air nozzle. Sistem quality checker dapat mendeteksi dan mengklasifikasi tiga kategori penempatan label yaitu accept, reject, dan no label dengan kecepatan konveyor yang berjalan 21,99 cm/s dan jarak 7cm dari kamera dan mempunyai rate acceptance sebesar 98,88%.
=====================================================================================================================================
Improvements made by PT PTI used the SBL (single bottom label) machine for in-line attachment to the VSN 03 production line. The SBL machine still did not have a checker attached to the labels. Therefore, the operator post then needs to do two jobs, namely to carry out a quality checker on the results of attaching the SBL machine and to enter the packaging into the dusat. The quality checker process has drawbacks, namely the occurrence of human errors caused by several factors including the accuracy and sharpness of the human eye that varies. In this research, a sorting machine system was created which was equipped with a detection system for attaching lip cream product labels using the YOLO (You Only Look Once) deep learning algorithm. This algorithm processes images in real time at 45 fps. YOLO detects objects using the unfield model where a single convolutional network predicts several bounding boxes as well as the class probabilities within these boxes simultaneously. The results of product detection and classification will be displayed on the monitor and then the products that are not good will be sorted by blowing them using an air nozzle. The quality checker system can detect and classify three categories of label placement, namely accept, reject, and no label with a conveyor speed of 21,99 cm/s and a distance of 7 cm from the camera and has an acceptance rate of 98,88%

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Kecacatan Label Produk lip cream , Quality Control, Deep Learning, You only Look Once, Sistem Sortir; lip cream Product Label Defects, Quality Control, Deep Learning, You only Look Once, Sorting System
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering
Divisions:	Faculty of Vocational > 36304-Automation Electronic Engineering
Depositing User:	Moch Deny Triatmaja
Date Deposited:	30 Aug 2023 07:27
Last Modified:	30 Aug 2023 07:27
URI:	http://repository.its.ac.id/id/eprint/102070

Actions (login required)

View Item