Search

Detecting road cracks is essential for inspecting and assessing the integrity of concrete pavement structures. Traditional image-based methods often require complex preprocessing to extract crack features, making them challenging when dealing with noisy concrete surfaces in diverse real-world scenarios, such as autonomous vehicle road detection. This study introduces an image-based crack detection approach that combines a Random Forest machine learning classifier with a deep convolutional neural network (CNN) to address these challenges. Three state-of-the-art models, namely MobileNet, InceptionV3, and Xception, were employed and trained using a dataset of 30,000 images to build an effective CNN. A systematic comparison of validation accuracy across various base learning rates identified a base learning rate of 0.001 as optimal, achieving a maximum validation accuracy of 99.97%. This optimal learning rate was then applied in the subsequent testing phase. The robustness and flexibility of the trained models were evaluated using 6,000 test photos, each with a resolution of 224 × 224 pixels, which were not part of the training or validation sets. The outstanding results, boasting a remarkable 99.95% accuracy, 99.95% precision, 99.94% recall, and a matching 99.94% F1 Score, unequivocally affirm the efficacy of the proposed technique in precisely identifying road fractures in photographs taken on real concrete surfaces. [ABSTRACT FROM AUTHOR]

As one of the spatial dimensions, orbital angular momentum (OAM) expands the channel capacity of underwater wireless optical communication (UWOC) system. However, the performance of the OAM-UWOC system is degraded due to the distortion of the OAM by oceanic turbulence. In this paper, we present a 7-layer convolutional neural network (CNN) and validate its ability to identify OAM patterns carried by Bessel Gaussian (BG) beams in oceanic turbulence channels. The trained CNN performs the task of recognizing OAM patterns with an accuracy of over 99% under weak-to-moderate oceanic turbulence at 100 m. Furthermore, in order to demonstrate the performance and robustness of the designed CNN, the effects of training sample number, epoch, turbulence intensity, balance parameter, transmission distance, type of training dataset, and coding method on the recognition ability of CNN are fully investigated. We hope that the method devised in our work will contribute to UWOC and facilitate its further development. [ABSTRACT FROM AUTHOR]