| Marine surveying and mapping is an important branch of surveying and mapping science.Modern marine surveying and mapping includes many technical fields such as marine geodetic surveying,seabed topographic surveying,hydrographic surveying and oceanographic remote sensing surveying.The appearance of ship-borne above and underwater integrated mapping system,multi-beam bathymetry system and laser scanning system has brought marine surveying and mapping into a new stage of development.Above and underwater terrain is the basis of marine economic development and maritime military activities.Traditional topographic surveying and mapping of islands(reefs),river banks,tidal flats,dock channels,etc.uses a combination of near-shore RTK and underwater single-beam sonar.Affected by factors such as terrain and environment,such methods have low efficiency and data collection density,and operators have high workload and certain risks.How to efficiently and accurately obtain topographic data on the water and underwater has become a research hotspot for marine surveying and mapping workers.Based on this requirement,this paper integrates laser scanner,multi-beam sounding sonar and positioning and attitude system to form a ship-borne integrated mapping system,which better solves the problem of above and underwater surveying and mapping,and is carried out in the subsequent data processing research.This paper mainly focuses on the point cloud data acquisition,calculation and data denoising and smoothing of lidar scanning system and multi-beam bathymetry system.The following researches are carried out:(1)Using the shipborne laser scanner,multi-beam echo sounder,GNSS positioning system and attitude meter,the autonomous construction and point cloud calculation of the integrated above and underwater system are realized.GPS and 1PPS are used to solve the time matching problem of multi-source sensor data;round-trip survey lines are laid out to obtain the placement error angle for system calibration,and finally the problem of spatial datum unification is solved through coordinate system transformation.(2)Various factors affecting the mapping accuracy of laser scanners and multi-beam sonar are systematically analyzed,and the evaluation methods for the accuracy of underwater and underwater point clouds are summarized.Aiming at the problem that it is difficult to carry out mapping work in complex waters such as wharfs and ports,the catamaran unmanned vessel is used to carry a small shallow water multi-beam sonar to realize data collection in complex waters below level 2 sea conditions.Through the lake test and sea test,the reliability of the system was verified,and the qualitative and quantitative evaluation of the mapping quality effect caused by the time matching and placement errors was carried out.(3)For traditional median filtering,it is difficult to deal with cluster outliers in the local range and small-scale outlier detection.This paper innovatively introduces two-step pre-screening to improve the robustness of median filtering.Point distance statistics are performed under the global point cloud,weights are calculated in layers in the local window,and thresholds are set to reduce the influence of noise on the median filter threshold calculation.Under different threshold settings,the detection rate of the improved median filter for noise is above 90%,which can save 60% of the time compared with manual processing.(4)For multi-beam point cloud data processing with with low signal-to-noise ratio,a combined filtering method based on noise classification is proposed.The noise points are divided into large-scale noise points far from the terrain main body and small-scale noise points close to the terrain main body,and the large-scale noise points are removed by radius filtering and statistical filtering;then the filtered data of large-scale noise points is used as the original data of smoothing filtering.Combined filtering combines the advantages of the two methods,effectively preserving the terrain features and the amount of effective data. |
PDF Full Text Request