A neural network-based methodology for pavement crack detection and classification

Kaseko, Mohamed S.; Ritchie, Stephen G.

doi:10.1016/0968-090x(93)90002-w

Cited by 145 publications

(45 citation statements)

References 2 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Some of the possible classifiers can be listed as Bayes classifier [15], k-nearest neighborhood classifier [16], multilayer feed-forward artificial neural networks [17], support vector machines [18], and etc. However, in most of the recent studies performed for crack detection, where several classifiers are compared, the accuracy of the results obtained by using neural network classifiers is shown to be higher than the other listed methods ( [8]; [9]; [14]; [19]). Thus, for this research, the neural network classifier is selected to be used for classification.…”

Section: Classificationmentioning

confidence: 99%

“…This prevents most of the current approaches from being used for crack quantification, since these methods are specifically developed for crack detection rather than quantification. For current approaches, it is required to maintain a constant focal length, resolution, or distance to the object in order to be able to extract crack dimensions ( [8]; [9]; rd International Symposium on Automation and Robotics in Construction (ISARC 2016) [10]; [11]; [12]; [13]). The damage detection method discussed in authors' previous research eliminates the requirement for prior knowledge on focal length, resolution, or distance to the investigated object, since all the required parameters for the defect detection are extracted from the point cloud automatically.…”

Section: Improvements For Crack and Spalling Detectionmentioning

confidence: 99%

See 1 more Smart Citation

Automated Classification of Detected Surface Damage from Point Clouds with Supervised Learning

Guldur¹,

Hajjar²

2016

Proceedings of the International Symposium on Automation and Robotics in Construction (IAARC)

View full text Add to dashboard Cite

-Recent advances in sensing technologies provide opportunity to utilize advanced data collection equipment such as laser scanners, high-resolution cameras, etc. for both short-and long-time monitoring of structures. Laser scanners, which are capable of collecting up-to two million points per second along with high-resolution images, are especially evolving rapidly and their usage for capturing and documenting the current condition of varying structural types is becoming increasingly feasible. The authors' previous research has focused on developing generalized algorithms for detecting surface damage on structures from captured laser scans and images. These algorithms are capable of automatically detecting surface damage on varying structural types as well as several construction materials since they use underlying surface geometry for performing damage detection. These damage types include concrete cracking, concrete spalling, steel delamination, steel section loss, bent members, and other misalignments. This paper investigates the use of supervised learning methods for determining whether a detected region represents an actual damage location. First, the input feature representations of the learning algorithms for each damage type are determined and the associated training sets, which are representative of real-world use of the algorithms, are prepared. Once the training and validation are completed, the learning algorithms are tested on the detected damage data, which is obtained by running the damage detection algorithms on both laser scans and texture-mapped images from a concrete test setup. Finally, the accuracy of these learning algorithms is investigated.

show abstract

Section: Classificationmentioning

confidence: 99%

Section: Improvements For Crack and Spalling Detectionmentioning

confidence: 99%

Automated Classification of Detected Surface Damage from Point Clouds with Supervised Learning

Guldur¹,

Hajjar²

2016

Proceedings of the International Symposium on Automation and Robotics in Construction (IAARC)

View full text Add to dashboard Cite

show abstract

“…The results contain less false detections but they are highly dependent on the choice of the parameters. Neuron networks-based methods have been proposed to alleviate the problems of the two first categories (Kaseko and Ritchie, 1993). However, they need a learning phase which is not well appropriate to the task.…”

Section: Detection Of Road Cracksmentioning

confidence: 99%

“…the crack, inside a known texture). (Bray et al, 2006;Chou et al, 1995;Kaseko and Ritchie, 1993;Ritchie et al, 1991) MULTI-SCALE (1990-2009) (Chambon et al, 2009;Delagnes and Barba, 1995;Fukuhara et al, 1990;Subirats et al, 2006;Zhou et al, 2005) …”

Section: Detection Of Road Cracksmentioning

confidence: 99%

ROAD CRACK EXTRACTION WITH ADAPTED FILTERING AND MARKOV MODEL-BASED SEGMENTATION - Introduction and Validation

Chambon

Gourraud

Moliard

et al. 2010

Proceedings of the International Conference on Computer Vision Theory and Applications

View full text Add to dashboard Cite

Abstract:The automatic detection of road cracks is important in a lot of countries to quantify the quality of road surfaces and to determine the national roads that have to be improved. Many methods have been proposed to automatically detect the defects of road surface and, in particular, cracks: with tools of mathematical morphology, neuron networks or multiscale filter. These last methods are the most appropriate ones and our work concerns the validation of a wavelet decomposition which is used as the initialisation of a segmentation based on Markovian modelling. Nowadays, there is no tool to compare and to evaluate precisely the peformances and the advantages of all the existing methods and to qualify the efficiency of a method compared to the state of the art. In consequence, the aim of this work is to validate our method and to describe how to set the parameters.

show abstract

“…From the outset these ANN applications have been used as support tools for management decision-making as they complement the already existing rule-based expert systems. Sundin and Braban-Ledoux [10] identify three principal areas for the application of neural networks to PMS: the first area involves estimating the current pavement condition [11,12,13], the second, predicting the future pavement condition [14] and the third, assessing the pavement needs and selecting the best maintenance actions [15,16]. In the estimation of current and also the prediction of future pavement condition, the neural network utilizes as inputs the different pavement characteristics and as targets, the pavement performance indicators such as ride quality or surface distress.…”

mentioning

confidence: 99%

Application of an ANN-based methodology for road surface condition identification on mining vehicles and roads

Ngwangwa

Heyns

2014

Journal of Terramechanics

View full text Add to dashboard Cite

 No vehicle models are employed for generation of training data. Vehicles are tested in their normal operating environments. Road profiles are crudely measured yet of satisfactory frequency content. The methodology is tested for different vehicle speeds and road profiles. AbstractAn artificial neural networks-based methodology for the identification of road surface condition was applied to two different vehicles in their normal operating environments at two mining sites.An ultra-heavy haul truck used for hauling operations in surface mining and a small utility underground mine vehicle were utilised in the current investigation. Unlike previous studies where numerical models were available and road surfaces were accurately profiled with profilometers, in this study, that was not the case in order to replicate the real mine road management situation. The results show that the methodology performed very well in reconstructing discrete faults such as bumps, depressions or potholes but, owing to the inevitable randomness of the testing conditions, these conditions could not fit the fine undulations present on the arbitrary random rough surface. These are better represented by the spectral displacement densities of the road surfaces. Accordingly, the proposed methodology can be applied to road condition identification in two ways: firstly, by detecting, locating and quantifying any existing discrete road faults/features, and secondly, by identifying the general level of the road's surface roughness.2

show abstract

A neural network-based methodology for pavement crack detection and classification

Cited by 145 publications

References 2 publications

Automated Classification of Detected Surface Damage from Point Clouds with Supervised Learning

Automated Classification of Detected Surface Damage from Point Clouds with Supervised Learning

ROAD CRACK EXTRACTION WITH ADAPTED FILTERING AND MARKOV MODEL-BASED SEGMENTATION - Introduction and Validation

Application of an ANN-based methodology for road surface condition identification on mining vehicles and roads

Contact Info

Product

Resources

About