Research On Pattern Recognition For Oil Stealing Detection Of Oil Pipelines

This dissertation focuses on the application of pattern recognition technology on stress wave based oil stealing remote monitoring system for the oil pipelines and the realization the whole system.A key problem to be solved in this system is to recognize the signal made by the stealing. This paper employs a statistical pattern recognition method of multi characteristic parameters to analyze the oil stealing signals and other signals propagated through the oil pipeline. After analyzing a great deal of field data, effective characteristic parameters are picked out. Through field tests, good classifying effect is achieved and the pattern recognition of signals in pipelines is well solved. By using the technologies of sensor, data acquisition, communication, multimedia and OOP (Object Oriented Programming), this system is developped to provide a real-time monitoring method for oil pipeline.This dissertation can be divided into two sections: the research of technology of pattern recognition of stress wave's signal and the realization of system's software. The former section includes the research background of pattern recognition, the process of signal recognition and the results of the experiments, while the later one discusses the design and realization of the system software.