Sparse representation of complex steerable pyramid for machine fault diagnosis by using non-contact video motion to replace conventional accelerometers

Research output: Journal Publications and Reviews (RGC: 21, 22, 62)21_Publication in refereed journalpeer-review

4 Scopus Citations
View graph of relations

Detail(s)

Original languageEnglish
Article number109104
Number of pages18
Journal / PublicationMeasurement
Volume175
Online published7 Feb 2021
Publication statusPublished - Apr 2021

Abstract

Structural health monitoring is vital for ensuring structural safety, avoiding property damage. Contact-type vibration measurement devices, such as accelerometers that provide high-spatial-resolution sensing capacity, cannot be installed on the inspected objects if their surface is hot or inaccessible. In contrast, non-contact-type vibration measurement devices, such as laser vibrometers, are relatively expensive and perform measurements sequentially, which could be time- and labor-intensive when the inspected areas are large. Recently, an alternative non-contact method has been proposed to measure vibration using a high-speed digital-video camera. Due to the continuous advancement in technology, these high-speed cameras are relatively low-cost and have simultaneous measurement capacity with high spatial resolution. To convert the video images captured by the camera to vibration displacement, the key is to use appropriate video signal processing method. For this purpose, phase-based analysis is often used. However, this analysis is time-consuming. Therefore, this paper presents a new processing method by combining the multiple-scale pyramid decomposition method and sparse representation technique. This method can considerably reduce the number of redundant steerable filters, improve the computation efficiency of phase-based video processing and replace the conventional mounted sensors. The experiments indicate that the results obtained by the proposed method for machine shaft fault diagnosis are comparable to those obtained using a physically mounted accelerometer. Hence, the proposed method provides a workable solution to detect abnormal vibrations generated by a structure/machine, with the advantages of no contact and low cost. In the future, this method could be used, for example, to measure the heart and breathing rates of elderly persons without a sensor placed on the body during sleeping.

Research Area(s)

  • Image processing, Machine Monitoring System, Sparse representation, Steerable pyramid, Video-based