Title :
A novel method for disturbance detection, localization and pattern recognition in signal images
Author :
Zadeh, Hossein Ghayoumi ; Janianpour, Siamak ; Haddadnia, Javad
Author_Institution :
Dept. of Electr. Eng., Sabzevar Tarbiat Moallem Univ., Sabzevar, Iran
Abstract :
A new approach to pattern recognition and classification of non-stationary power signal is presented in this paper. In the proposed work visual localization and detection of non-stationary power signals are achieved using Image processing and its pattern is recognized and classified by MLP neural network algorithm. Also disturbance localization and classification of the signal, is done once with image processing and one more time with neural network and the results are compared with each other. In MLP neural network method we tried to reduce the disturbance localization errors of image processing method. Various non-stationary power signals are processed through image processing stage to generate property charts of the signal for extracting relevant features for pattern classification. The extracted features are clustered using MLP Neural Network algorithm to refine the cluster centers.
Keywords :
feature extraction; image classification; multilayer perceptrons; MLP neural network algorithm; cluster centers; disturbance classification; disturbance localization; feature extraction; image processing; nonstationary power signal detection; pattern classification; pattern recognition; property charts; visual localization; Accuracy; Artificial neural networks; Feature extraction; Image edge detection; Noise; Pixel; Image processing; MLP neural network algorithm; Non-stationary power signal; Pattern Recognition;
Conference_Titel :
Intelligent Engineering Systems (INES), 2011 15th IEEE International Conference on
Conference_Location :
Poprad
Print_ISBN :
978-1-4244-8954-1
DOI :
10.1109/INES.2011.5954755