Research On Feature Extraction Of Pipeline Defects Based On Mathematical Morphology

With the acceleration of economic construction, the amount of oil and gas being used is increasing day after day. The resource crisis such as oil shortage on land becomes more and more serious. Because of the rich oil resources, the exploitation of ocean petroleum resource is extensively concerned. Pipeline transportation has many advantages in long distance transportation of oil, gas and other liquid resources. But concerning that lots of the pipelines in our country have been used too many years, there must be many defects in those pipelines. So aiming at the aging status of the pipelines and frequent accident, the timely detection, life evaluation and maintenance of the pipelines being used are very necessary. So pipeline defects’detection is of great significance.Through the technology and processing of pipeline magnetic flux leakage detect, and the research on the data of different pipeline defects from pipeline magnetic flux leakage detect, the phenomenon that different defections of pipelines have different magnetic flux leakage signals is found.Aiming at the problem such as the pipeline magnetic flux leakage image noise, edge blur and low contrast, a pretreatments method to pipeline magnetic flux leakage detect image is proposed. By using of noise filter algorithm and image enhancement algorithm, the quality of pipeline magnetic flux leakage image is improved. And by the using of image adaptive binary defect extraction method, the area of pipeline defect is extracted. Through the experimental simulation, the pretreatments method to pipeline magnetic flux leakage image is proved effective.Based on the pretreatments to the pipeline magnetic flux leakage image and aiming at the problem of the boundary determination of pipeline defects, the method that based on mathematical morphology to deal with pipeline magnetic flux leakage image is put forward. Mathematical morphology is a method that can treat the image characters by observing the structural elements’movements in the image. The pipeline defects’boundary is determined by using mathematical morphology. And then, the chain code information of pipeline defects is marked and saved by the using of mark algorithm and boundary tracking algorithm. Through the simulation, the purposes of the boundary determination and boundary information extraction are achieved.And then, aiming at the problem of the description and recognition the shape of pipeline defect, some parameters that can describe the shape of defects are chosen and the formulation of these parameters are given. Based on the boundary chain code information of pipeline defects and boundary tracking scanning and calibration conversion formulas, the area, circumference, long shaft length and short shaft length, dispersion degree and eccentricity degree of every defect are calculated. The purposes of detection feature extraction are accomplished. And the result is satisfying.Finally, based on the work of the thesis, the research and work in the future are prospected.