Identifikasi Ketersediaan Ruang Parkir Menggunakan Faster R-Cnn Dengan Resnet-50

Reswara, Muhammad Andhika (2025) Identifikasi Ketersediaan Ruang Parkir Menggunakan Faster R-Cnn Dengan Resnet-50. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5002211086-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (38MB) | Request a copy

Abstract

Seiring bertumbuhnya populasi penduduk, banyak permasalahan yang mengiringinya seperti meningkatnya polusi dan padatnya lalu lintas. Permasalahan tersebut dikarenakan pertumbuhan jumlah kendaraan bermotor yang terus meningkat. Akibat dari banyaknya kendaraan ini seringkali dirasakan ketika pengendara hendak parkir di suatu tempat yang ramai seperti pusat perbelanjaan. Dengan munculnya permasalahan tersebut, perlu adanya sistem yang dapat mempermudah pengendara dalam menemukan tempat parkirnya. Penelitian ini bertujuan untuk mengidentifikasi ketersediaan tempat parkir menggunakan Faster R-CNN dengan ResNet-50 sebagai backbone nya. Proses penelitian ini mencakup tahap pre-processing, training, validasi, pengujian, pembahasan, dan penarikan kesimpulan. Model ini dilatih dan diuji menggunakan data primer dan data sekunder. Hasil dari pengujian model ini mendapatkan nilai mAP mencapai 95,57% untuk cuaca cerah, 97,87% untuk kondisi cuaca berawan, dan 93,40% untuk kondisi cuaca hujan. Selain itu, didapatkan bahwa performa klasifikasi yang diperoleh dengan metrik evaluasi mencapai sebesar 94,16% untuk precision, 96,12% untuk recall, 91,58% untuk accuracy, dan 94,96% untuk F1-score.
========================================================================================================================================
As the population grows, there are many problems that come with it such as increased pollution and traffic congestion. These problems are caused by the increasing number of motorised vehicles. The result of this large number of vehicles is often felt when motorists want to park in a crowded place such as a shopping centre. With the emergence of these problems, there is a need for a system that can facilitate motorists in finding parking spaces. This research aims to identify the availability of parking spaces using Faster R-CNN with ResNet-50 as its backbone. This research process includes pre-processing, training, validation, testing, discussion, and conclusion. The model was trained and tested using primary and secondary data. The results of testing this model obtained mAP values reaching 95.57% for sunny weather, 97.87% for cloudy weather conditions, and 93.40% for rainy weather conditions. In addition, it was found that the classification performance obtained by evaluation metrics reached 94.16% for precision, 96.12% for recall, 91.58% for accuracy, and 94.96% for F1-score.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Faster R-CNN, ResNet-50, Parkir, Deteksi, Performansi, Faster R-CNN, ResNet-50, Parking, Detection, Performance
Subjects:	Q Science > QA Mathematics > QA336 Artificial Intelligence Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science)
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Mathematics > 44201-(S1) Undergraduate Thesis
Depositing User:	Muhammad Andhika Reswara
Date Deposited:	22 Jul 2025 08:26
Last Modified:	22 Jul 2025 08:26
URI:	http://repository.its.ac.id/id/eprint/120645

Actions (login required)

View Item