J-Icon : Jurnal Komputer dan Informatika
Vol 9 No 2 (2021): Oktober 2021

DETEKSI KEBERADAAN RAMBU PERINTAH DI JALAN MENGGUNAKAN EUCLIDEAN COLOR FILTERING

Rizal Pebianto (Universitas Nusa Cendana)
Adriana Fanggidae (Unknown)
Yulianto Triwahyuadi Polly (Unknown)



Article Info

Publish Date
26 Oct 2021

Abstract

Command signs are one of the types of traffic signs that we often encounter when driving on the road and are used to express some commands while driving. A distinctive feature of command signs is that they have a blue background color. Driver negligence in paying attention to the presence of command signs on the road can result in accidents. There needs to be a system that can detect command signs on the road as a solution. The command sign detection process uses the Euclidean Color Filtering method to filter colors and BLOB to look for interconnected pixels. The RGB color center values ​​(10,110,200) and the radius range [90, 150] are used in the color filter stage with Euclidean Color Filtering. The system was tested on 150 images of primary data and 25 images of secondary data. Primary data collection was carried out in the morning, afternoon, and evening with a distance of 5–10 meters and 10–15 meters. The system's average accuracy of command sign detection in testing primary data is 90.67%, and secondary data is 76%, so the average accuracy of the whole system is 88.5%. System failure in detecting the presence of command signs on the road is due to insufficient or excessive lighting conditions, signs blocked by other objects, angles that are too tilted, and signs adjoining other things with the same color or close to the background color.

Copyrights © 2021






Journal Info

Abbrev

jicon

Publisher

Subject

Computer Science & IT

Description

J-ICON : Jurnal Komputer dan Informatika focuses on the areas of computer sciences, artificial intelligence and expert systems, machine learning, information technology and computation, internet of things, mobile e-business, e-commerce, business intelligence, intelligent decision support systems, ...