Image of PENGOLAHAN DATA CITRA PADA SISTEM DETEKSI DAN MAPPING JALAN BERLUBANG MENGGUNAKAN DEEP LEARNING DAN GPS DI WILAYAH SUMATERA SELATAN
Bookmark Share

Skripsi

PENGOLAHAN DATA CITRA PADA SISTEM DETEKSI DAN MAPPING JALAN BERLUBANG MENGGUNAKAN DEEP LEARNING DAN GPS DI WILAYAH SUMATERA SELATAN

According to data from the Ministry of Public Works and Public Housing (PUPR) in 2022, South Sumatra is one of the regions with a high rate of road damage in Indonesia, including numerous potholes. To address this issue, an automatic pothole detection system using deep learning and GPS-based mapping was developed. This system involves data preprocessing and mapping to create a representative dataset for training the deep learning model to recognize the patterns and characteristics of potholes. The process includes data mining and image processing techniques such as object detection, segmentation, and pattern recognition. This study focuses on collecting datasets in South Sumatra. Data collection will be conducted using a ground vehicle equipped with a GoPro Hero 8 action camera for real-time video capture and GPS for pothole mapping. The collected data must accurately reflect pothole conditions, with varying camera angles to ensure precise representation. The collected data is used to train the deep learning model to identify various patterns and characteristics of potholes. This research successfully implemented image data processing in a pothole detection and mapping system using deep learning and GPS, achieving average error UIQI values 0,074, indicating high similarity images and GPS Neo-6M demonstrated good accuracy with an average distance error below 1 meter. Data preprocessing significantly improved the performance of YOLOv8 and MaskRCNN models, proving that preprocessing helps the model achieve better performance across various evaluation metrics.

Availability
#
Central Library (Reference) T1515712024
T151571
Available but not for loan - Not for Loan
Detail Information
Series Title
-
Call Number
T1515712024
Publisher
Indralaya : Prodi Teknik Elektro, Fakultas Teknik Universitas Sriwijaya.,
Collation
x, 73 hlm.; ilus.; 29 cm
Language
Indonesia
ISBN/ISSN
-
Classification
006.370 7
Content Type
Text
Media Type
unmediated
Carrier Type
-
Edition
-
Subject(s)
Pengolahan Citra Digital
Prodi Teknik Elektro
Specific Detail Info
-
Statement of Responsibility
Other version/related
File Attachment
Comments
You must be logged in to post a comment