MLS Point Cloud Intensity Correction And Application Research For Shield Tunneling

Shield tunnels are a key force in transportation,shortening the distance of traffic roads and saving ground space.Due to the complex underground environment and the high volume of passengers that the tunnel receives,the consequences of a safety accident are unimaginable.This paper takes shield tunnel data as the research object,analyzes tunnel intensity information,proposes an intensity correction model,and performs tunnel feature extraction and point cloud classification based on the corrected intensity data and the generated tunnel surface intensity images to achieve the extraction of pipe wall data and detection of water seepage areas,and accomplishes the following work and research results:(1)Data acquisition and preprocessing of shield tunnel:For data acquisition of shield tunnel,a set of mobile laser scanning equipment is designed and assembled,and a reasonable scanning plan is formulated with full consideration of various factors to conduct data acquisition on the surface of shield tunnel.Since there are various noises in the data,this paper de-noises the collected data according to the types of noises.Since the cylindrical space of the tunnel is not conducive to observation and analysis,it is developed into a two-dimensional plane and then converted into an intensity image.(2)Exploration and correction of influence mechanism of shield tunnel intensity:Laser intensity value contains reflection characteristics of tunnel surface,which can be used for tunnel feature extraction and disease detection.However,the laser intensity data is affected by distance and incidence Angle,so it needs correction to extract the target surface characteristics.The influence of distance and incident Angle on intensity value is explored by using homogenous target region,and the precise influence mechanism of distance and incident Angle on intensity value is obtained.According to the influence mechanism,a unified correction model was established to complete the correction of the strength value.The intensity correction is carried out about 50m point cloud data collected independently from 350m to 400m of a shield tunnel in Shanghai.The experiment shows that the mean intensity of the measured data is 216.755,and the standard deviation is 36.906,while the mean intensity after correction is 171.981,and the standard deviation is 11.378,and the standard deviation decreases significantly.Coefficient of variation decreased by 61%from 17.03%to 6.62%,and coefficient of variation ratio was0.39,indicating a small degree of data dispersion.(3)Application of strength information in tunnel feature extraction and point cloud classification:integration of geometric information and strength information to extract pipe wall data;Water seepage detection was carried out on the tunnel point cloud to obtain the permeable water area.Take about 50m tunnel point cloud data at 550m to 600m as an example,the original tunnel data has 30,701,388 point cloud data,and 24,644,005 point cloud data of tunnel data can be obtained by eliminating the adjunct based on geometric information,effectively eliminating 6,057,383(19.72%)point cloud data.The tunnel wall data were obtained according to the point cloud intensity.The number of point clouds on the tunnel wall was23,040,067,and 1,603,938(5.22%)supplementary point clouds were screened out.The total number of supplementary point clouds in this section was 7,661,321(24.95%).It shows that the pipe wall extraction method combining point cloud geometry data and strength data can correctly separate the pipe wall data from the accessory point cloud,and effectively eliminate the tunnel wall accessories,such as pipelines,cables,transformer boxes and bolt holes.Taking about 30m tunnel point cloud data at 610m to 640m as an example,the tunnel water leakage area was extracted and the area statistics were carried out.In this section of the tunnel,there are a total of 45 water leakage areas.The water leakage detection algorithm based on strength data proposed in this paper correctly detected 42 water leakage areas,and basically all the water leakage areas were successfully identified,indicating that the tunnel disease identification algorithm based on feature removal proposed in this paper has good applicability.According to the leakage area,the actual leakage area of the tunnel is 4.17m~2,and the leakage area accounts for 1.05%of the total area.