TINJAUAN SISTEMATIS KOMPARASI PCI DAN VISI KOMPUTER UNTUK EVALUASI KERUSAKAN JALAN
This study aims to conduct a comparative analysis of the conventional Pavement Condition Index (PCI) method and a computer vision-based approach in evaluating road pavement damage in Indonesia. The method used was a Systematic Literature Review (SLR) based on the PRISMA 2020 guidelines. Data were collected through a structured search of the Scopus and Google Scholar databases, identifying 27 selected studies that met the inclusion criteria. The results show that PCI has high accuracy and is internationally recognized for evaluating 20 types of damage, but it requires significant time, cost, and manpower and involves subjectivity. Conversely, computer vision using deep learning algorithms can provide fast, objective, and scalable results with 76–95% accuracy, although it is still limited to specific types of damage and is influenced by environmental factors. This study recommends a hybrid approach that integrates both methods, along with the use of low-cost technologies, such as smartphones and GIS-based UAVs, to support more efficient road infrastructure maintenance in Indonesia.