Optical Flow-Based Detection of Gas Leaks from Pipelines Using Multibeam Water Column Images
Optical Flow-Based Detection of Gas Leaks from Pipelines Using Multibeam Water Column Images
复制标题
使用多束水柱图像进行基于光流的管道气体泄漏检测
DOI：
10.3390/rs12010119
复制
发表时间：
2020-01-01
期刊：
影响因子：
5
通讯作者：
White, Paul R.
中科院分区：
文献类型：
--
作者：
Xu, Chao;Wu, Mingxing;White, Paul R.
近年来，大多数多波束回声测深仪（MBES）已能够收集水柱图像（WCI）数据，同时进行海底地形测量，提供了有效的数据源，气体泄漏检测。然而，可以存在系统的（例如，旁瓣干扰）或图像中的自然干扰，这可能对气体泄漏的自动检测带来挑战。在本文中，我们设计了两种数据处理方案来估计运动速度的Farneback光流原理的基础上，根据类型的WCI，包括时间-角度和深度跨轨道图像。此外，通过将估计的运动速度与图像像素的幅度相结合，使用多个判决阈值来消除干扰，例如海底，水柱中的非气体后向散射等。为了验证所提出的方法的有效性，我们模拟了中国吉林省松花湖和水池中的管道泄漏场景，并使用HT 300 PA MBES（由哈尔滨工程大学开发，其工作频率为300 kHz）在静态和动态条件下收集声学数据。结果表明，该方法可以自动检测水下泄漏气体，两种数据处理方案具有相似的检测性能。
In recent years, most multibeam echo sounders (MBESs) have been able to collect water column image (WCI) data while performing seabed topography measurements, providing effective data sources for gas-leakage detection. However, there can be systematic (e.g., sidelobe interference) or natural disturbances in the images, which may introduce challenges for automatic detection of gas leaks. In this paper, we design two data-processing schemes to estimate motion velocities based on the Farneback optical flow principle according to types of WCIs, including time-angle and depth-across track images. Moreover, by combining the estimated motion velocities with the amplitudes of the image pixels, several decision thresholds are used to eliminate interferences, such as the seabed, non-gas backscatters in the water column, etc. To verify the effectiveness of the proposed method, we simulated the scenarios of pipeline leakage in a pool and the Songhua Lake, Jilin Province, China, and used a HT300 PA MBES (it was developed by Harbin Engineering University and its operating frequency is 300 kHz) to collect acoustic data in static and dynamic conditions. The results show that the proposed method can automatically detect underwater leaking gases, and both data-processing schemes have similar detection performance.