Extreme cross-correlation-based laser backscattering detection of turbulence-coupled bubble wakes

The optical characteristics of marine vehicle bubble wakes demonstrate significant dynamic properties in complex flow fields, with optical information coupled to various turbulent flow field parameters, exhibiting features such as wide dynamic range and non-stationary background. This poses severe challenges to traditional fixed-threshold detection methods. To address issues including low signal-to-noise ratio in laser wake detection signals due to bubble and thermal turbulence, limited continuous effective detection cycles, and low wake feature capture probability, this study innovatively integrates a dynamic threshold optimization mechanism with cross-correlation extreme value detection technology. A extreme correlation-based method for detecting backward optical signals of bubble wakes is proposed. Moreover, a collaborative processing architecture is designed, integrating a dynamic autocorrelation background fluctuation suppression algorithm and a target signal enhancement algorithm utilizing cross-correlation extreme values. Furthermore, a dynamic threshold interval is established based on high-correlation characteristics of wake-free background signals. The sliding-window cross-correlation extreme value method is employed to separate target signals from background noise while enhancing target features, thus overcoming limitations of traditional methods in non-stationary fluid environments and resolving continuous multi-cycle wake detection challenges. Additionally, a multi-turbulent-field-coupled vehicle wake simulation platform is constructed to verify the detection method’s accuracy and reliability under turbulent interference. Experimental results indicate a significant improvement in the average signal-to-background ratio (SBR) by 5.01 dB. At a 95 % detection confidence level, bubble wake feature capture rates in turbulent environments increase from 58.3 % to 86 %, continuous effective detection time extends to 2.72 times that of conventional methods, and detection sensitivity (via reduced response time) improves by an order of magnitude. Consequently, this work provides a potential application method for extracting optical features of bubble wakes in complex fluid environments.