Defect Classification in NDT Applications using Time Frequency Features, LDA, and a KNN Classifier

Deriche, Mohamed
May 2014
Australian Journal of Basic & Applied Sciences;May2014, Vol. 8 Issue 7, p485
Academic Journal
In this paper, we develop a new approach for detecting defects in steel. We show that given the time varying nature of the acquired signals, we need to use nonstationay signal processing approaches. In particular, we focus on the performance of different time-frequency distributions (TFD). We show that the extraction of robust features from such TF distributions can lead to excellent classification accuracy of different defects independent of the specific classifier used. To evaluate the performance of our system, we compare our results with those obtained using the conventional 71 features traditionally used in benchmarking algorithms. Our experimental results using artificial and real defects show that the use of TF features provides an excellent characterization of defects in steel. Classification is carried used a dimension reduction technique, namely Linear Discriminant Analysis (LDA), followed by a KNN classifier. Additionally, we show that some specific TFDs can perform better than others in steel defect detection. In particular, the Gabor transform is shown to yield the best classification accuracy among the different time-frequency distributions considered in this study.


Related Articles

  • Novel decision-making technique for damage diagnosis. Gelman, L; Murray, B; Patel, T H; Thomson, A // Insight: Non-Destructive Testing & Condition Monitoring;Aug2013, Vol. 55 Issue 8, p428 

    The article presents a study which focuses on the implementation, and validation of novel anomaly detection technique based on the k-nearest neighbours (kNN) method for damage diagnosis for bearings. It mentions that proposed method consists of several stages including data clustering,...

  • DRBF and IRBF Based Face Recognition and Extraction of Facial Expressions from the Blur Image. Jayashree, M.; Deepa, D.; Rubhashree, M. // Australian Journal of Basic & Applied Sciences;Mar2014, Vol. 8 Issue 3, p61 

    Background: Facial expression is the natural for human beings to show their emotions and motivations. When blurring comes to picture, face recognition becomes difficult. The main factors that make this problem challenging are image degradation due to blur and the appearance variations due to...

  • Influence of the number of topologically interacting neighbors on swarm dynamics. Shang, Yilun; Bouffanais, Roland // Scientific Reports;2/28/2014, p1 

    Recent empirical and theoretical works on collective behaviors based on a topological interaction are beginning to offer some explanations as for the physical reasons behind the selection of a particular number of nearest neighbors locally affecting each individual's dynamics. Recently, flocking...

  • Pulmonary Acoustic Signal Classification using Autoregressive Coefficients and k-Nearest Neighbor. Palaniappan, Rajkumar; Sundaraj, Kenneth; Sundaraj, Sebastian; Huliraj, N.; Revadi, S. S.; Archana, B. // Applied Mechanics & Materials;2014, Issue 591, p211 

    Pulmonary acoustic signals provide important information of the condition of the respiratory system. It can be used to assist medical professionals as an alternative diagnosis tool. In this paper, we intend to discriminate between normal (without any pathological condition), Airway Obstruction...

  • DC Motor Fault Analysis with the Use of Acoustic Signals, Coiflet Wavelet Transform, and K-Nearest Neighbor Classifier. Glowacz, Adam // Archives of Acoustics;Sep2015, Vol. 40 Issue 3, p321 

    This paper focuses on testing the monitoring system of the Direct Current motor. This system gives the possibility of diagnosing various types of failures by means of analysis of acoustic signals. The applied method is based on a study of acoustic signals generated by the DC motor. A study plan...

  • Faulty measurements impact on wireless local area network positioning performance. Morales, Jafet Aaron; Akopian, David; Agaian, Sos // IET Radar, Sonar & Navigation;Jun2015, Vol. 9 Issue 5, p501 

    Indoor positioning methods based on wireless local area network (WLAN) signal measurements have gained popularity because of high localisation accuracy. These methods use radio-maps obtained from wireless signal measurement surveys on location grids. Measurement sets from various WLAN access...

  • Application of Vector Quantization for Audio Retrieval. Vaidya, Shruti; Shah, Kamal // International Journal of Computer Applications;Feb2014, Vol. 88, p23 

    Due to the progress of the unlimited data storage capabilities and the proliferation use of the Internet, information retrieval systems encountered a large interest. Much of this data is in different forms from various sources. So, it becomes important to develop the necessary technologies for...

  • Vehicle Type Recognition in Sensor Networks Using Improved Time Encoded Signal Processing Algorithm. Yan Wang; Xi Wu; Xiaohua Li; Jiliu Zhou // Mathematical Problems in Engineering;2014, p1 

    Vehicle type recognition is a demanding application of wireless sensor networks (WSN). In many cases, sensor nodes detect and recognize vehicles from their acoustic or seismic signals using wavelet based or spectral feature extraction methods. Such methods, while providing convincing results,...

  • Advanced signal processing and condition monitoring. Martin, N. // Insight: Non-Destructive Testing & Condition Monitoring;Aug2007, Vol. 49 Issue 8, p459 

    When physical models are of a high complexity, a signal processing approach is helpful for providing accurate information about a system and its failures. The session at this Conference entitled ‘Advanced Signal Processing and Condition Monitoring’ contains papers who propose new...


Read the Article


Sorry, but this item is not currently available from your library.

Try another library?
Sign out of this library

Other Topics