Submarine pipeline and leakage point detection method

Abstract

The invention provides a submarine pipeline and leakage point detection method comprising the following steps: step 1, shooting by using an underwater camera to obtain an underwater image, then carrying out enhancement processing on the underwater image, denoising the image by using a Gaussian filtering method, enabling the image to be clearer by using histogram equalization, and enhancing the image brightness by using a low-illumination image enhancement algorithm; step 2, establishing an underwater environment submarine pipeline optical image detection data set, and carrying out a detectionexperiment on underwater pipeline image and a video by using a YOLOv3 algorithm; and step 3, making an underwater environment submarine pipeline optical image segmentation data set, carrying out imagesegmentation on the underwater pipeline and a leakage target, and fusing and improving a latest instance segmentation algorithm MASK R-CNN algorithm and a YOLACT algorithm to realize an optimal imagesegmentation effect. According to the invention, the MASK R-CNN and YOLACT algorithms are fused and improved to carry out image segmentation on submarine pipelines and leakage points. The submarine pipeline and leakage point detection method can meet the requirement of submarine pipeline and leakage point real-time detection, and is high in detection accuracy.

Description

technical field [0001] The invention relates to a detection method, in particular to a detection method of a submarine pipeline and a leakage point, and belongs to the field of target detection and image segmentation. Background technique [0002] The purpose of target recognition is to find out whether there is a proposed category in a given image (such as whether there are pipelines and leaks in the underwater submarine pipeline image), and if there is, output the specific location of the category and use the bounding box to frame it. The current challenge of target recognition is still how to improve the accuracy and recognition speed of the algorithm, and how to adapt to the changing environment. Especially in the underwater environment, the complex underwater environment will make it more difficult for the recognition algorithm to adapt. Therefore, we need to carry out continuous research on the target recognition algorithm to make it more adaptable to the underwater en...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide a detection method for submarine pipelines and leakage points that can overcome the interference of complex underwater environments

Images