CD Skripsi

Sistem Pendeteksi Bola Dan Gawang Dengan Algoritma Convolutional Neural Network Pada Robot Krsbi Menggunakan Kamera Omnidirectional

---

ABSTRACT
Kontes Robot Sepak Bola Indonesia-Beroda is a competition held to enhance students's knowledge and creativity in robotics. Competing robots must be able to dribble the ball and score goals autonomously. The HSV color filtering method is commonly used for object detection; however, it has limitations in handling variations in lighting intensity. This study proposes the implementation of a Convolutional Neural Network (CNN) algorithm based on You Only Look Once (YOLO) to improve the accuracy and stability of ball and goal detection using an omnidirectional camera. By using a dataset of 1,125 images of balls and goals, divided into 80% for training and 20% for validation, a model was obtained with an accuracy of 93.8% and an F1-Score reaching 1.00 at a confidence level of 0.883. Furthermore, after conducting detection tests to compare the performance of YOLO with HSV, the ball detection accuracy using HSV was found to be 28% in the morning, 64% in the afternoon, and 71% at night. In contrast, YOLO achieved an accuracy of 85% in the morning, 92% in the afternoon, and 100% at night. For goal detection, HSV achieved 50% accuracy in the afternoon and 66% at night, while the YOLO model successfully reached 100% accuracy under all lighting conditions.
Keywords : KRSBI-Beroda, YOLO, HSV, ball detection, goalpost detection, omnidirectiolan camera

---

Ketersediaan

Informasi Detail

Judul Seri

-

No. Panggil

2007113753

Penerbit

Pekanbaru : Universitas Riau – F.TEKNIK – INFORMATIKA., 2025

Deskripsi Fisik

-

Bahasa

Indonesia

ISBN/ISSN

-

Klasifikasi

2007113753

Tipe Isi

-

Tipe Media

-

Tipe Pembawa

-

Edisi

-

Subjek

TEKNIK INFORMATIKA

Info Detail Spesifik

-

Pernyataan Tanggungjawab

daus

Versi lain/terkait

Lampiran Berkas

Komentar

---

Anda harus masuk sebelum memberikan komentar