Research On The Key Techniques Of Above And Underwater Integrated Mapping System

Marine surveying and mapping is a discipline that studies the collection,processing,representation,management,and application of geospatial information such as oceans,rivers,lakes,and adjacent land areas.It is an important branch of surveying and mapping and the basis of marine scientific research and development and utilization activities.The integrated mapping of above and underwater is a great significance as a branch of marine mapping.Traditional island and reef,river bank mapping are usually carried out with RTK in conjunction with single-beam sonar.This method is affected by constraints such as topography,geomorphology,and environmental conditions,and usually has low mapping efficiency.How to acquire both the above and underwater topography data efficiently has become a difficult problem in the field of surveying and mapping.Based on this problem,in the times of implementing a marine power,the combination of a shipborne Li DAR scanning system and a multi-beam sonar system to form an integrated above and underwater mapping system has great innovation and practicality.This paper focuses on the following researches on data acquisition and classification based on the integrated above and underwater mapping system consisting of a shipborne Li DAR scanning system and a multi-beam sonar system:(1)This article introduces in detail principle and composition of the shipborne Li DAR scanning system and multi-beam sonar system.The resolution of Pulse Per Second PPS solves the problem of high-precision time synchronization of multiple sensors and the rigid connection of multiple sensors solves the problem of spatial registration.This article put forward a set of practical solutions for integrated mapping of above and underwater.(2)Analyze the shipborne Li DAR scan data and navigation positioning and orientation system data,and jointly solve to generate 3D point cloud data,and use the calibration field to eliminate the placement angle deviation of the shipborne Li DAR data,and verify the shipborne Li DAR scanning system.The accuracy reaches the centimeter level,which meets the 1: 1000 precision map drawing.The water depth data correction processing such as gross error elimination,tide level correction,sound speed correction and transducer installation deviation compensation are performed on the multi-beam sonar system to achieve the optimal state of the multi-beam sonar system.(3)Analyze the sources of noise in the above and underwater integrated mapping system,introduce several commonly used denoising algorithms,and perform radius filtering automatically denoising on the data of the shipborne Li DAR scanning system and the multi-beam sonar system,instead of the traditional human-computer interaction denoising method,greatly improved the efficiency,can save more than 70% time of the internal processing,and provides technical support for efficient mapping.(4)In this paper,the improved Point Net network and the current high-precision semantic segmentation of large scenes Rand LA-Net neural network,are used to segment water and underwater integrated point cloud data.The original point cloud data and threedimensional features are used as inputs to achieve the respectively.The overall classification accuracy of 86.57% and 87.81% verifies the accuracy and universality of the method,and it can be applied to the semantic segmentation of point cloud data in most shipborne waterborne and underwater integrated mapping scenarios.This article uses efficient deep learning and artificial neural network algorithms to perform semantic segmentation of point cloud data for above and underwater integration and point cloud data for most Li DAR scenes,complete classification work,and generates DEM and 3D reconstruction.This work provides an effective method.