Irhamsyah, Muhammad Zahid Riefqi and Prof. Dr.Eng. Agus Naba, S.Si., M.T and Dewi Anggraeni, S.Si., M.SI (2024) Implementasi Sistem Deteksi Cacat Botol Susu Berbasis Embedded System Jetson Agx Orin Dan Bahasa Pemrograman Python. Sarjana thesis, Universitas Brawijaya.
Abstract
Inspeksi kecacatan adalah salah satu upaya dalam mendeteksi kecacatan dalam hasil produksi massal. Transisi industri 4.0 mengajak pihak industri untuk menerapkan Internet of Things terutama Artificial Intelligence (AI). Integrasi AI di Industri memudahkan prediksi keputusan dalam mendeteksi kualitas botol susu yang dipasarkan. Penelitian ini bertujuan agar mengembangkan sistem deteksi cacat botol susu berbasis embedded system menggunakan Jetson AGX Orin serta mendapatkan kinerja sistem yang optimal. Integrasi dua AI (Support Vector Machine dan Random Forest) sekaligus digunakan pada penelitian ini digunakan agar memprediksi objek pada kondisi spesifik. Support Vector Machine digunakan sebagai mendeteksi kehilangan huruf label dan Random Forest untuk mendeteksi penyok pada botol susu. Penelitian ini memberikan hasil bahwa Jetson AGX Orin dan kamera dapat diterapkan sebagai sistem deteksi kecacatan botol susu secara realtime. Hasil deteksi terekam sebagai gambar tangkapan kamera dan gambar yang telah diproses pada channel warna grayscale. Integrasi Support Vector Machine memiliki akurasi model AI 1.00 dan Random Forest memiliki akurasi 0.95. Kinerja selama sistem deteksi berjalan mendapatkan maksimum konsumsi yang tinggi pada GPU dan cukup rendah pada CPU. Total konsumsi pada CPU sebesar 18.1% dan total konsumsi GPU adalah 49%. Sistem deteksi yang dijalankan menggunakan memori RAM sebesar 4.1 GB.
English Abstract
In order to prevent defective products exports worldwide, the industry should implement product defect inspection. The transition into industry 4.0 invites industries to implement internet of things notably Artificial Intelligence (AI). Integration of AI in industry can make easier prediction in detecting quality of milk bottle that exported worldwide. This research is intended to develop a detection system of defects from milk bottle based on embedded system through the usage of Jetson AGX Orin and able to achieve the optimal performance of the system. The integration of two AI models useful to predict specific condition. Support Vector Machine is useful to detect defects for lack of text labels while Random Forest can be use as dent detection of a milk bottle. This research concludes that Jetson AGX Orin and camera can be used as a detection system of defects on milk bottle in realtime. The results will be added as an image on the system as camera capture and image that has been processed in grayscale channel. The integration of Support Vector Machine has model accuracy at 1.00 and Random Forest has model accuracy at 0.95. When the system is running, System detection needed high amount of GPU consumption while also needed low amount of CPU consumption. Total consumption of CPU is 18.1 % and the total consumption of GPU is 49%. When detection system is running, the total RAM memory usage is 4.1 GB.
Item Type: | Thesis (Sarjana) |
---|---|
Identification Number: | 052409 |
Uncontrolled Keywords: | Model AI, sistem deteksi, prediksi |
Divisions: | Fakultas Matematika dan Ilmu Pengetahuan Alam > Instrumentasi Fisika |
Depositing User: | Unnamed user with username nova |
Date Deposited: | 22 Aug 2024 04:17 |
Last Modified: | 22 Aug 2024 04:17 |
URI: | http://repository.ub.ac.id/id/eprint/223683 |
![]() |
Text (DALAM MASA EMBARGO)
Muhammad Zahid Riefqi Irhamsyah.pdf Restricted to Registered users only Download (3MB) |
Actions (login required)
![]() |
View Item |