In view of the complex time-varing signal pattern recognition, the method of water flooded layer recognition based on wavelet packet transform and support vector machine had been proposed in this paper. According to the multi-resolution characteristics of wavelet packet analysis, the denoising for signals and the extraction of useful signals had been presented. The sample space had been mapped into the high dimensional feature space by employing the support vector machine. At the same time, the classification of signals had been also achieved by constructing the optimal separating hyperplane. The experiment showed that this method had the better classification performance and robustness.
展开▼
