Implementasi Deep Learning dengan Metode Convolutional Neural Network (CNN) pada Arm Robot berbasis kamera untuk deteksi warna objek

Salsabila, Lailiana (2025) Implementasi Deep Learning dengan Metode Convolutional Neural Network (CNN) pada Arm Robot berbasis kamera untuk deteksi warna objek. Sarjana thesis, UIN Sunan Gunung Djati Bandung.

[img]
Preview
Text (COVER)
1_cover.pdf

Download (291kB) | Preview
[img]
Preview
Text (ABSTRAK)
2_abstrak.pdf

Download (221kB) | Preview
[img]
Preview
Text (LEMBAR PERNYATAAN)
3_skbebasplagiarism.pdf

Download (413kB) | Preview
[img]
Preview
Text (DAFTAR ISI)
4_daftarisi.pdf

Download (275kB) | Preview
[img]
Preview
Text (BAB I)
5_bab1.pdf

Download (238kB) | Preview
[img] Text (BAB II)
6_bab2.pdf
Restricted to Registered users only

Download (726kB) | Request a copy
[img] Text (BAB III)
7_bab3.pdf
Restricted to Registered users only

Download (518kB) | Request a copy
[img] Text (BAB IV)
8_bab4.pdf
Restricted to Registered users only

Download (1MB) | Request a copy
[img] Text (BAB V)
9_bab5.pdf
Restricted to Registered users only

Download (217kB) | Request a copy
[img] Text (DAFTAR PUSTAKA)
10_daftarpustaka.pdf
Restricted to Registered users only

Download (208kB) | Request a copy

Abstract

INDONESIA: Telah dikembangkan sebuah Arm Robot dengan 4 Degree of Freedom (DoF) yang mampu mendeteksi dan mengklasifikasikan objek berdasarkan warna menggunakan kamera (webcam) dan metode Convolutional Neural Network (CNN) dengan model VGG16. Sistem ini dirancang untuk mengenali warna objek yang ditangkap oleh ka- mera, kemudian memindahkannya ke kotak yang sesuai dengan warnanya. Pemodelan dilakukan menggunakan bahasa pemrograman Python. Pengujian dilakukan terhadap empat kelas warna (merah, biru, kuning, hitam) serta variasi intensitas cahaya (17 lux, 47 lux, 70 lux, dan 88 lux). Sistem mampu mendeteksi warna dengan baik pada selu- ruh variasi, kecuali pada intensitas cahaya 17 lux dan 47 lux yang menghasilkan satu kegagalan deteksi akibat pencahayaan yang terlalu rendah. Uji coba juga dilakukan pa- da variasi latar belakang dan bentuk objek (segitiga, bulat dan belah ketupat), dimana seluruh warna target tetap dapat dikenali dengan baik kecuali 1 percobaan pada belah ketupat dengan intensitas cahaya sebesar 17 lux. Hasil menunjukkan bahwa bentuk dan latar belakang objek tidak memberikan pengaruh yang signifikan terhadap proses klasifikasi warna oleh sistem. English: A 4-Degree of Freedom (DoF) Arm Robot has been developed, capable of detecting and classifying objects based on color using a webcam and the Convolutional Neural Network (CNN) method with the VGG16 model. This system is designed to recognize the color of objects captured by the camera and then move them to a box corresponding to their color. The modeling was carried out using the Python programming language. Testing was conducted on four color classes (red, blue, yellow, black) under varying light intensities (17 lux, 47 lux, 70 lux, and 88 lux). The system was able to detect co- lors well under all lighting variations, except at 17 lux and 47 lux, where one detection failure occurred due to insufficient lighting. Trials were also conducted with varying backgrounds and object shapes (triangle, circle, and rhombus), where all target colors were still successfully recognized, except for one instance with a rhombus-shaped ob- ject under 17 lux lighting. The results show that object shape and background do not significantly affect the system’s color classification process.

Item Type: Thesis (Sarjana)
Uncontrolled Keywords: Deep Learning; Convolutional Neural Network; deteksi warna; arm robot; kamera; VGG16 Deep Learning; Convolutional Neural Network; color detection; robotic arm; camera; VGG16.
Subjects: Physics > Research and Statistical Methods of Physics
Physics > Instrumentation of Physics
Light, Infrared and Ultraviolet Phenomena > Physical Optics
Divisions: Fakultas Sains dan Teknologi > Program Studi Fisika
Depositing User: Lailiana Salsabila
Date Deposited: 07 Jan 2026 06:54
Last Modified: 07 Jan 2026 06:54
URI: https://digilib.uinsgd.ac.id/id/eprint/127201

Actions (login required)

View Item View Item