The Research Of Submarine Pipeline Inspection Technology Based On Side-scan Sonar System

Pipeline laid on the seabed plays an important role in oil transportation.Due to the influence of the tides cyclical fluctuation the submarine near pipeline willbe eroded, moreover, the long-term erosion can cause the pipeline with the states ofnudity, dangling, and displacement. All kinds of status will affect the normaloperation of the submarine pipeline. Currently, detection the state of submarinepipeline relys mainly on the side-scan sonar collecting pipe of acoustic images andreading the images by people, but this method has some shortcomings such as lowefficiency and poor scientific. To solve these problems, the dissertation usedintelligence algorithm of BP network and sparse-representation to achieve the stateof the submarine pipeline identification and classification, the main research work asfollows:Firstly, the dissertation used DGPS technology and side-scan sonar to collectside scan sonar data of submarine pipeline hangzhou bay submarine pipeline, andanalyzed the reasons of various state for pipeline, which provides scientificexperimental data for further research.Secondly, in the research of side-scan sonar image recognition, the capturedimages from software usually regard as data source, but the image on the displaysoftware can’t response the size of the sampling data. So in the dissertation,analysing the data of side-scan sonar(XTF format), and extracting sampling datawhich can response to the information of seabed form the data of format, andconverting sampling data to grey value matrix which can be used for imageclassification. Then, using Visual studio2010to develop software that used todisplay the sampling data scroll, and the software is convenient experimentalcontrast research.Finally, using BP neural network and Sparse represent the model of thesubmarine pipeline state classification recognition algorithm. Then, compared theexperimental results of two models, Elaborate two kinds of classification recognitionalgorithm of their advantages, provide direction for the state of submarine pipelinedetection in the future.