Abstract: First, in order to reconstruct three-phase power quality disturbance signals, the Fourier basis is chosen as the over-complete dictionary; Gaussian random matrices as the observation matrix; and orthogonal pursuit algorithm as the reconstruction algorithm. Then, the energy eigenvalue of gray level cooccurrence matrix texture features and the probability of the occurrence of grey value are used to classify the disturbance signals. The experiment results show that the proposed method can not only realize simultaneous compression and reconstruction of three-phase electric signals and accurate classification of such signals, but also reconstruct the original signal samples, reducing the data volume of signal compression and sampling.

Key words: information processing, power quality, compressed sensing(CS), grayscale, gray level cooccurrence matrix