Detection and assessment of cracks in civil engineering structures such as roads, bridges, dams and pipelines are crucial tasks for maintaining the safety and cost-effectiveness of those concrete structures. With the recent advances in machine learning, the development of ANN- and CNN-based algorithms has become a popular approach for the automated detection and identification of concrete cracks. However, most of the proposed models are trained on images taken in ideal conditions and are only capable of achieving high accuracy when applied to the concrete images devoid of irregular illumination conditions, shadows, shading, blemishes, etc. An overview of challenges related to the automatic detection of concrete cracks in the presence of shadows is presented in this paper. In particular, difficulties associated with the application of deep learning-based methods for the classification of concrete images with shadows are demonstrated. Moreover, the limitations of the shadow removal techniques for the improvement of the crack detection accuracy are discussed.
Deep learning-based methods, especially convolutional neural networks, have been developed to automatically process the images of concrete surfaces for crack identification tasks. Although deep learning-based methods claim very high accuracy, they often ignore the complexity of the image collection process. Real-world images are often impacted by complex illumination conditions, shadows, the randomness of crack shapes and sizes, blemishes, and concrete spall. Published literature and available shadow databases are oriented towards images taken in laboratory conditions. In this paper, we explore the complexity of image classification for concrete crack detection in the presence of demanding illumination conditions. Challenges associated with the application of deep learning-based methods for detecting concrete cracks in the presence of shadows are elaborated on in this paper. Novel shadow augmentation techniques are developed to increase the accuracy of automatic detection of concrete cracks.
In this study, two categories of persons with normal and high ABP are subjected to the bicycle stress test (9 persons with normal ABP and 10 persons with high ABP). All persons are physically active men but not professional sportsmen. The mean and the standard deviation of age is 41.11 ± 10.21 years; height 178.88 ± 0.071 m; weight 80.53 ± 10.01 kg; body mass index 25.10 ± 2.06 kg/m2. Machine learning algorithms are employed to build a set of rules for the classification of the performance during the stress test. The heart rate, the JT interval, and the blood pressure readings are observed during the load and the recovery phases of the exercise. Although it is obvious that the two groups of persons will behave differently throughout the bicycle stress test, with this novel study, we are able to detect subtle variations in the rate at which these changes occur. This paper proves that these differences are measurable and substantial to detect subtle differences in the self-organization of the human cardiovascular system. It is shown that the data collected during the load phase of the stress test plays a more significant role than the data collected during the recovery phase. The data collected from the two groups of persons are approximated by Gaussian distribution. The introduced classification algorithm based on the statistical analysis and the triangle coordinate system helps to determine whether the reaction of the cardiovascular system of a new candidate is more pronounced by an increased heart rate or an increased blood pressure during the stress test. The developed approach produces valuable information about the self-organization of human cardiovascular system during a physical exercise.
A novel feature extraction technique is presented in this paper. The term H-rankgram is coined here due to the similarity in concept with other feature extraction methods like spectrogram. The H-rankgram is two-dimensional feature pattern which shows the change in algebraic complexity (measured using ranks of Hankel matrices) of a given signal at a different scale in time (estimated using phase space reconstruction with different time lags). In general feature extraction techniques could be classified into two broad types: time domain and frequency domain. The proposed technique fits into the first one. The proof of concept for the technique to detect changes in the signal was explained and an effect of additive noise was tested. Application of the technique was demonstrated to classify ECG signals for healthy and ill patients. The results shows that Myocardial infarction is detected most accurately although there is high imbalance in classification accuracy between the classes.
Convolutional neural networks have been created as deep learning-based approaches to automatically analyze photographs of concrete surfaces for crack diagnosis applications. Although deep learning-based systems assert to have extremely high accuracy, they frequently overlook how difficult it is to acquire images. Complex lighting situations, shadows, the irrationality of crack forms and widths, imperfections, and concrete spall frequently have an influence on real-world photos. The focus of the published research and accessible shadow databases is on photographs shot in controlled laboratory settings. In this research, we investigate the challenging underwater optical effects settings and the complexity of image classification for concrete crack detection. This research elaborates on difficulties encountered when using deep learning-based techniques to identify concrete cracks when optical effects are present. To improve the precision of automatically detecting concrete cracks on underwater surfaces, new optical effect augmentation techniques have been developed.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.