Pipeline multi-sensor data integration based on D-S evidence theory

In order to ensure the accuracy and credibility of pipeline detection, multiple detection techniques i.e. multi-sensor are commonly used to detect the same part, but the detected data, miscellaneous and redundant, needs to be integrated and processed. By using D-S (Dempster-Shafer) evidence theory, this paper sets up a process model of multi-sensor data integration for pipeline, and presents the process of data integration. Based on detective credibility of various sensors to different defects, detected data is integrated by using D-S evidence theory, and types of defect are determined according to D-S evidence theory. The results show that by using D-S to integrate data in the testing process, and keeping high credibility data, a clearer decision can be made, abandoning interferential data, saving storage space, good for data storage, and achieving intelligent detection in long distance.