Article / članak 15

Integration of non-destructive testing and machine learning in pavement management systems

Integracija nerazornih ispitivanja i strojnog učenja u sustav gospodarenja kolnicima

authors / autori
Iva Brkić, Josipa Domitrović

DOI
https://doi.org/10.5592/CO/PhDSym.2026.15

keywords / ključne riječi
pavement management, homogeneous sections, non-destructive testing, ground penetrating radar (GPR), falling weight deflectometer (FWD), machine learning, pavement segmentation
sustav gospodarenja kolnicima, homogene dionice, nerazorne metode ispitivanja, georadar, deflektometar s padajućim teretom, strojno učenje, segmentacija kolnika

page range in publication / raspon stanica
125 - 132

publish date / datum publikacije
24/9/2026

abstract / sažetak
In pavement management systems, the identification of homogeneous sections is important for reliable pavement condition assessment and maintenance planning. This paper presents an overview of homogeneous section identification, the role of non-destructive testing methods, with a particular focus on ground penetrating radar and the falling weight deflectometer, as well as the potential application of machine learning. The integration of these approaches can improve the reliability of pavement segmentation and provide better support for decision-making.
U sustavu gospodarenja kolnicima određivanje homogenih dionica je važan preduvjet za pouzdanu procjenu stanja kolnika i planiranje mjera održavanja. U radu je dan pregled područja određivanja homogenih dionica, uloge nerazornih metoda ispitivanja, s naglaskom na georadar i deflektometar s padajućim teretom, te mogućnosti primjene strojnog učenja. Zaključno, integracija tih pristupa omogućuje pouzdaniju segmentaciju kolnika i kvalitetniju podršku odlučivanju.