Independent Detecting And Recognizing Pipeline On The Seabed

The subject stems from the project of "One Autonomous Underwater Vehicle" . That AUV cruises in the fixed date, which can independently find and track pipeline on the seabed. It not only can observe the state of pipeline and the circumstance near the pipeline, but also can warn the coming fault. It is important to independently detect and recognize pipeline on the seabed for the AUV finding and tracking pipeline. The system of detecting and recognizing pipeline includes image collection , image pre-processing, image segmentation , feature extraction and pattern recognition.Firstly, considering the characteristics of underwater image, we have done the research on the image segmentation. A new rapid edge detection arithmetic base on the classic Kirsch arithmetic operators is used in the thesis, which improves the speed and effect of edge detection. We presented a novel threshold algorithm based on maximum entropy principle, in which a PSO algorithm was used to search the. optimal threshold in the two-dimensional gray space. Statistical results of experiments showed that the new algorithm presented in this paper can find better solutions with little cost, and this method was feasible and effective.Finally, the paper has built the system of recognizing pipeline on the seabed. The eigenvectors of pipeline's image are got through the feature extraction and selection, which are the six feature vectors. Experiment showed that the six feature vectors had better clustering effect. BP neural net are applied to recognize pipeline's image. This method is tested with six feature vectors. Theoretical and experimental results showed that the way extracting different features of different patterns could distinguish different patterns effectively. In conclusion, it is shown that the algorithms and approached providing in the thesis are effective and instructive to go on the in-depth study of independently detecting and recognizing pipeline on the seabed.