In this study, the application of four classification techniques for computer vision–based pavement crack detection systems was investigated. The classification methods—artificial neural network (ANN), decision tree, k–nearest neighbor, and adaptive neuro-fuzzy inference system (ANFIS)—were selected on the basis of the complexity and clarity of their procedures. These methods were evaluated for ( a) prediction performance, ( b) computation time, ( c) stability of results for highly imbalanced data sets, ( d) stability of the classifiers’ performance for pavements in different deterioration stages, and ( e) interpretability of results and clarity of the procedure. According to the results, the ANN and ANFIS methods not only provide superior performance but also are more flexible and compatible for the crack detection application. The ANFIS method is called a “white-box classifier,” and the inferred knowledge from its membership functions can be used to characterize the imagery properties of detected image components.
Automated pavement distress recognition is a key step in smart infrastructure assessment. Advances in deep learning and computer vision have improved the automated recognition of pavement distresses in road surface images. This task remains challenging due to the high variation of defects in shapes and sizes, demanding a better incorporation of contextual information into deep networks. In this paper, we show that an attention-based multi-scale convolutional neural network (A+MCNN) improves the automated classification of common distress and non-distress objects in pavement images by (i) encoding contextual information through multi-scale input tiles and (ii) employing a mid-fusion approach with an attention module for heterogeneous image contexts from different input scales. A+MCNN is trained and tested with four distress classes (crack, crack seal, patch, pothole), five non-distress classes (joint, marker, manhole cover, curbing, shoulder), and two pavement classes (asphalt, concrete). A+MCNN is compared with four deep classifiers that are widely used in transportation applications and a generic CNN classifier (as the control model). The results show that A+MCNN consistently outperforms the baselines by 1∼26% on average in terms of the F-score. A comprehensive discussion is also presented regarding how these classifiers perform differently on different road objects, which has been rarely addressed in the existing literature.
The reliable detection of relatively small changes in the characteristics of monitored systems, which simultaneously involve nonlinear phenomena as well as uncertain parameters, is a challenging problem whose resolution is crucial to the development of practical structural health monitoring methodologies slated for use with complex physical systems. This paper reports the results of a comprehensive experimental study involving an adaptive nonlinear component (an actively controlled magnetorheological damper) that was used to investigate the representation and propagation of uncertainties in a probabilistic format that provides a convenient means for reliable detection of small changes in uncertain nonlinear systems. In experimental studies of the MR damper, the uncertainty of the system characteristics was precisely controlled with known input-current statistics. A total of 4000 tests were performed, and the MR damper was identified using the restoring force method with both orthogonal and non-orthogonal basis functions. The identification results show that the identified coefficients involving orthogonal basis functions have several desirable features that are ideal for condition assessment purposes when dealing with complex nonlinear systems whose underlying physics is not amenable to easy modeling: (1) no a priori knowledge of the systems characteristics is required; (2) the orthogonal coefficients are statistically unbiased with respect to model complexity; and (3) the distributions of the orthogonal coefficients can be reliably used to detect changes in uncertain nonlinear systems, even if a reduced-order model is used in the identification process.
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.