Application Of Underwater Linear Objects Processing Algorithm In Submarine Pipeline Inspection Robot

The mysterious ocean is a chest containing boundless treasure, although we haven’t stopped for the exploration and pursuit of it, but we always failed to see how the ocean really is. The emergence of underwater robot provides a powerful tool for us to explore this mysterious field.With the increasing of offshore oil and gas fields, the submarine pipeline has formed the crisscrossed network. At the same time, its security concerns have also increased. The key factors influencing the security of the submarine pipeline is corrosion, so it has great significance for the research of corrosion detection methods. Traditional detection methods are mostly random testing, which are not comprehensive to test the submarine pipeline, and high cost, so the research for new methods of submarine pipeline detection is necessary. Underwater robots, as the most advanced detection equipment, play an important role in the field of submarine pipeline detection. Due to the corrosion of submarine pipeline, the electric field around submarine pipeline will produce distortion. So this paper adopts the ROV as submarine pipeline inspection robot, equipped with a magnetic detection system, positioning system, potential measurement system, and video capture system, through the way of non-contact detection to get the corrosion of submarine pipeline.The software, as an important part of submarine pipeline potential detection system, affects the level of corrosion detection accuracy. This paper describes the PC software development process in detail, and presents the integral design of hardware and software, and mainly introduces the establishment of communication protocol and software architecture design.Although the light attenuation and scattering is more serious in the water, which results the videos and images are not clear, but as the most comprehensive and abundant information, image is one of the key points we need further research. It has great significance for the research of submarine pipeline detection and corrosion. In this paper, the underwater linear object processing algorithm is studied. We have proposed a rapid detection method of lines based on contour information, and applied it in submarine pipeline inspection robot. First of all, the collected underwater images are smoothed and enhanced to weaken the noise, get high quality images. Then reduce the image size by using wavelet transform and get the LL subband, and detect it by Canny to get the binary image of edge, namely contour information. Finally, detect it by Hough, and get the existence of submarine pipeline in the image.The experimental results show that the proposed rapid detection method of lines based on contour information can quickly extract the linear information from underwater image, and meet the real-time requirements. Its application in submarine pipeline lays the foundation for automatic recognition and processing of submarine pipeline detection.