PROTOTYPE ALAT PENGHITUNG OBAT OTOMATIS MENGGUNAKAN CITRA VISUAL UNTUK KEBUTUHAN STOCK OPNAME BERBASIS IoT

MUHAMMAD NURDIN PRAKASA, . (2026) PROTOTYPE ALAT PENGHITUNG OBAT OTOMATIS MENGGUNAKAN CITRA VISUAL UNTUK KEBUTUHAN STOCK OPNAME BERBASIS IoT. Sarjana thesis, UNIVERSITAS NEGERI JAKARTA.

	Text Cover.pdf Download (3MB)
	Text BAB 1.pdf Download (1MB)
	Text BAB 2.pdf Restricted to Registered users only Download (3MB) | Request a copy
	Text BAB 3.pdf Restricted to Registered users only Download (4MB) | Request a copy
	Text BAB 4.pdf Restricted to Registered users only Download (3MB) | Request a copy
	Text BAB 5.pdf Restricted to Registered users only Download (826kB) | Request a copy
	Text DAFTAR PUSTAKA.pdf Download (1MB)
	Text LAMPIRAN.pdf Restricted to Registered users only Download (1MB) | Request a copy

Abstract

Proses stock opname obat merupakan kegiatan penting dalam pengelolaan persediaan obat di apotek dan fasilitas kesehatan. Namun, proses penghitungan stok obat masih banyak dilakukan secara manual sehingga membutuhkan waktu yang lama, rentan terhadap kesalahan perhitungan (human error), serta kurang efisien dalam pengelolaan data inventaris. Penelitian ini bertujuan untuk merancang dan membangun sistem stock opname obat otomatis berbasis computer vision dan Internet of Things (IoT) yang mampu melakukan penghitungan obat secara otomatis, menyimpan hasil penghitungan ke dalam file Microsoft Excel, serta mengirimkan data ke website secara real-time. Metode penelitian yang digunakan adalah metode rekayasa teknik yang meliputi tahap identifikasi masalah, studi literatur, perancangan sistem, pengembangan perangkat keras dan perangkat lunak, implementasi, serta pengujian sistem. Sistem dibangun menggunakan Raspberry Pi sebagai pusat kendali, kamera sebagai perangkat akuisisi citra, motor servo sebagai mekanisme pengguncang tray obat, serta antarmuka Graphical User Interface (GUI) berbasis Python. Metode deteksi objek yang digunakan adalah YOLOv8 Segmentation yang dilatih menggunakan dataset publik Roboflow sebanyak 6.028 citra yang terdiri dari 4.746 data pelatihan, 847 data validasi, dan 435 data pengujian. Hasil deteksi kemudian disimpan ke dalam file Microsoft Excel dan dikirim ke website melalui API Ingest yang dibangun menggunakan framework Flask dan database SQLite. Berdasarkan hasil pengujian, sistem mampu melakukan penghitungan obat secara otomatis dengan tingkat akurasi yang tinggi pada berbagai variasi jumlah obat. Pengujian pencahayaan menunjukkan bahwa kondisi pencahayaan normal sebesar 239 lux menghasilkan performa deteksi terbaik dibandingkan kondisi redup 29 lux dan terang 350 lux. Pengujian kecepatan menunjukkan bahwa sistem mampu menyelesaikan proses stock opname lebih cepat dibandingkan metode manual. Selain itu, mekanisme pengguncangan tray menggunakan motor servo terbukti mampu mengurangi tumpukan obat sehingga meningkatkan keberhasilan deteksi objek. Sistem juga berhasil menyimpan dan mengirimkan data hasil stock opname ke website secara real-time tanpa mengalami kegagalan komunikasi. Hasil penelitian menunjukkan bahwa sistem yang dikembangkan mampu meningkatkan efisiensi proses stock opname obat, mengurangi potensi kesalahan perhitungan, serta mendukung digitalisasi pengelolaan inventaris obat melalui integrasi teknologi computer vision, kecerdasan buatan, dan IoT. ***** Medicine stock-taking (stock opname) is an essential activity in managing pharmaceutical inventories in pharmacies and healthcare facilities. However, inventory counting is still commonly performed manually, requiring considerable time, being prone to human error, and resulting in inefficient inventory management. This study aims to design and develop an automated drug stock-taking system based on computer vision and the Internet of Things (IoT) that can automatically count medicines, store counting results in Microsoft Excel files, and transmit data to a web-based system in real time. The research employed an engineering method consisting of problem identification, literature review, system design, hardware and software development, implementation, and system testing. The developed system utilizes a Raspberry Pi as the main controller, a camera for image acquisition, a servo motor as a tray-shaking mechanism, and a Python-based Graphical User Interface (GUI). The object detection method applied in this study is YOLOv8 Segmentation, which was trained using a public Roboflow dataset containing 6,028 images, consisting of 4,746 training images, 847 validation images, and 435 testing images. The detection results are automatically stored in Microsoft Excel files and transmitted to a web server through an API Ingest service developed using the Flask framework and SQLite database. The testing results indicate that the system is capable of automatically counting medicines with a high level of accuracy across various quantities of drugs. Lighting condition tests showed that normal illumination at 239 lux provided the best detection performance compared to dim lighting at 29 lux and bright lighting at 350 lux. Speed testing demonstrated that the system completed the stock-taking process faster than conventional manual methods. Furthermore, the tray-shaking mechanism driven by a servo motor successfully reduced drug overlapping, thereby improving object detection performance. The system also successfully stored and transmitted stock-taking data to the website in real time without communication failures. The results demonstrate that the proposed system can improve the efficiency of pharmaceutical stock-taking processes, reduce counting errors, and support the digitalization of drug inventory management through the integration of computer vision, artificial intelligence, and IoT technologies.

Item Type:	Thesis (Sarjana)
Additional Information:	1). Taryudi, Ph.D; 2). Rizki Pratama Putra, S.T., M.T.
Subjects:	Teknologi dan Ilmu Terapan > Teknik Mesin, Mekanika Teknik Teknologi dan Ilmu Terapan > Teknik Elektronika Teknologi dan Ilmu Terapan > Teknik Komputer
Divisions:	FT > D IV Teknologi Rekayasa Otomasi
Depositing User:	Muhammad Nurdin Prakasa .
Date Deposited:	03 Aug 2026 04:08
Last Modified:	03 Aug 2026 04:08
URI:	http://repository.unj.ac.id/id/eprint/67266

Actions (login required)

View Item