Wavelet-based anomaly detection on digital signals
Ömer Aydın, Melek Kurnaz · 2017
Since today's information and telecommunication devices are very complex, it is time consuming task to find anomalies on digital signals which happen very rarely and not easy to recreate by testing. In this work, a new anomaly detection method is proposed which includes wavelet transform, k-means clustering and morphological operators such as closing and dilation to detect anomalies on digital signals. After wavelet decomposition was performed on the digital signals which recorded by the oscilloscope and then transferred to the computer, unsupervised learning and mathematical morphology have been applied. Test results show that the proposed method achieves high-detection rates.