No abstract

Introduction.Early defect illumination (cracks, chips, etc.) in the high traffic load sections enables to reduce the risk under emergency conditions. Various photographic and video monitoring techniques are used in the pavement managing system. Manual evaluation and analysis of the data obtained may take unacceptably long time. Thus, it is necessary to improve the conditional assessment schemes of the monitor objects through the autovision.Materials and Methods.The authors have proposed a model of a deep convolution neural network for identifying defects on the road pavement images. The model is implemented as an optimized version of the most popular, at this time, fully convolution neural networks (FCNN). The teaching selection design and a two-stage network learning process considering the specifics of the problem being solved are shown. Keras and TensorFlow frameworks were used for the software implementation of the proposed architecture.Research Results.The application of the proposed architecture is effective even under the conditions of a limited amount of the source data. Fine precision is observed. The model can be used in various segmentation tasks. According to the metrics, FCNN shows the following defect identification results: IoU - 0.3488, Dice - 0.7381.Discussion and Conclusions.The results can be used in the monitoring, modeling and forecasting process of the road pavement wear.

No abstract

The first part of this work is devoted to the location of defects in a coated bar and the identification of their geometrical parameters. Using the methods of finite element modeling, ultrasonic non-destructive testing and machine learning technologies (artificial neural networks), the inverse problem of mechanics has been solved. A finite element model of ultrasonic wave propagation in a bar with a coating and an internal defect is constructed. Compared with previous works, the model used PML (Perfectly Matched Layer) structures, which suppress multiple reflections of the probe ultrasound pulse inside the bar and prevent signal noise. Based on the conducted numerical calculations of the finite element model, a data set was constructed. It contains the geometrical parameters of the defect and the corresponding amplitude-time characteristic of the ultrasonic signal. The architecture of a direct propagation neural network has been developed. The neural network was trained on the basis of previously processed data. As a result, on the basis of ultrasound data obtained from the outer surface of the bar, it is possible to restore the values of such defect parameters as depth, length and thickness. At the second stage, analytical-numerical technology for studying the stress intensity factor (SIF) at the crack tip is described using the example of the problem of a longitudinal internal crack of finite length located in an elastic strip reinforced with a thin flexible coating. The solution to this problem is based on the method of integral transformations, which made it possible to reduce it to a singular integral equation of the first kind with a Cauchy kernel, which is solved by the collocation method in the form of expansion in Chebyshev polynomials with a factor that explicitly takes into account a feature in the vicinity of the crack vertices. The latter allows you to directly find the SIF and evaluate the effect on it of various combinations of geometric and physical parameters of the problem.

Design principles of a novel Multifunctional Operation Station (MOS) using Augmented Reality (AR) technology (MOSAR) are proposed in this paper. AR-based design allows more ergonomic remote instrument control in real time in contrast to classical instrument-centered interfaces. Another advantage is its hierarchical software structure including multiple programming interpreters. The MOSAR approach is illustrated with a remote surgical operating station that controls intelligent surgical instruments. The implementation of the Operation Station (MOS) is based on the multiplatform open-source library Tcl/Tk, and an AR extension has been developed on a Unity platform, using Vuforia SDK.