Extraction And Location Of Plume Based On Multi-beam Water Column Data

The marine cold spring is rich in mineral resources,and the submarine cold spring and related research have attracted much attention because of its potential energy factors and environmental benefits,and its outstanding physical feature lies in the bubble plume formed by a large number of bubbles above the cold spring.this phenomenon provides a new idea for the exploration of marine cold springs.The research on plume sensitive area identification,plume target extraction and plume leakage source location is of great significance to the quantitative estimation of cold spring resources and the exploration of marine resources in China.Multi-beam sonar has the characteristics of wide detection range and high efficiency.Its water column imaging is an effective means to study the release and upward migration of marine free gas.Plume bubbles can be recorded in the water column image in the form of backscattering intensity.For this reason,this paper studies the identification and location of plume flow based on multi-beam water column data.The main contents are as follows:(1)The system composition and imaging principle of multi-beam sonar are analyzed and summarized,the construction of multi-beam water column image is studied in detail,and a set of multi-beam water column data processing method is given.It mainly includes:original data decoding,ray tracing,position reduction,image generation,data interpolation,grayscale transformation,pseudo-color enhancement.Through the above processing,a more complete water column image is obtained.(2)In view of the huge amount of water column image data and the low efficiency of plume target water column image detection,in order to improve the detection efficiency and reliability,a comprehensive detection method was studied.By determining the relationship between the acoustic parameters and geological properties of the seafloor sediment in the plume leakage area,the seafloor sediment is classified automatically,and then the twodimensional plane recognition of the plume sensitive area is realized.Combined with the method of single beam heading profile,the determination of plume target Ping water column image is realized.(3)Aiming at the phenomenon of sidelobe effect and instrument self-noise reducing the image quality of water column image,by analyzing the characteristics of plume target and background noise in water column image,the relationship between the statistical characteristics of bubble plume flow and the corresponding back-scattering intensity of multi-beam water column image is established.Threshold segmentation,k-means clustering,fuzzy c-means clustering,kernel-based fuzzy c-means clustering and morphological constraints are used to separate plume targets from background noise.With the help of the three-dimensional target detection function of multi-beam sonar "navigational" observation,a three-dimensional plume filter is constructed,and then the three-dimensional extraction of plume targets is realized.(4)Aiming at the phenomenon that the plume information in water column data is incomplete due to the influence of ocean environmental noise and flow field,on the basis of plume extraction,the bubble rising path template is constructed,and the path matching method based on improved ICP algorithm is studied to restore the incomplete plume path template in the leachate area,so as to realize the precise location of the submarine leakage source of plume flow.