Intelligent Graph Convolutional Neural Network for Road Crack Detection
Peer reviewed, Journal article
Accepted version
Permanent lenke
https://hdl.handle.net/11250/3086723Utgivelsesdato
2022Metadata
Vis full innførselSamlinger
- Publikasjoner fra CRIStin - SINTEF AS [5850]
- SINTEF Digital [2523]
Originalversjon
IEEE transactions on intelligent transportation systems (Print). 2022, 24 (8), 8475-8482. 10.1109/TITS.2022.3215538Sammendrag
This paper presents a novel intelligent system based on graph convolutional neural networks to study road crack detection in intelligent transportation systems. The visual features of the input images are first computed using the well-known Scale-Invariant Feature Transform (SIFT) extraction algorithm. Then, a correlation between SIFT features of similar images is analyzed and a series of graphs are generated. The graphs are trained on a graph convolutional neural network, and a hyper-optimization algorithm is developed to supervise the training process. A case study of road crack detection data is analyzed. The results show a clear superiority of the proposed framework over state-of-the-art solutions. In fact, the precision of the proposed solution exceeds 70%, while the precision of the baseline methods does not exceed 60%.