Study on Monitoring System for Pipeline Security Based on Multi-Sensor Data Fusion Technique

Abstract

AbstractBased on multi-sensor data fusion technique, we designed monitoring system for pipeline security. The system acquires early warning signals by many sensors and processing modules when pipes are threatened. The non-stationary signal analysis method based on empirical mode decomposition is used to process the signals. The normalized kurtosis extracted from the decomposed results is composed of the feature vectors. Because there were separate sensors gathering modules in the monitoring system, we used the multi-sensor data fusion technique to fuse these individual recognition results for improving recognition rate. With software development platform, LabVIEW virtual instrumentation, this obtained the final decision. With the test, the system can achieve better detection performance than a system using the single sensor, so as to form a more complete, reliable pipeline real time detecting conclusion.KeywordsPipelineMulti-sensor dataFusion techniqueLabVIEW virtual instrumentation