MODELLING THE DEPENDENCE OF POINT CLOUD DENSITY ON SHOOTING SPEED

This paper discusses a numerical method for estimating the dependence of the final quality, as the density of the point cloud per square metre of road surface, on the speed of a vehicle using mobile laser scanning (MLS) equipment. The emphasis on this type of scanning material is due to the prevalence of MLS for surveying linear motorways. With the introduction of MLS, the road industry has received a powerful tool for detailed analysis of the condition of road surfaces and infrastructure. MLS technologies allow you to create high-density point clouds and high-resolution panoramic images that accurately display the objects of the road network and elements of traffic management (TSODD). The MLS data not only helps in assessing the quality of the coating and identifying defects but is also used for modeling and positioning objects on highways, which significantly improves the accuracy of road infrastructure management. The study proposes the author's method of data analysis and automation tools for its implementation: scripts written in Python and sqlscripts for PostgreSQL + PostGIS. The materials for the analysis are the data of mobile laser scanning of public roads of regional purpose in St. Petersburg, taken in different speed modes with the total length of 3.1 km with the number of lanes from four to six. The point clouds underwent primary classification with extraction of analysed ground points using the CSF filter plugin in the open-source software CloudCompare. Along this axis, auxiliary polygons (between adjacent scanning stations in both directions from the axis of movement) are subdivided, which are used to estimate the density of points at a distance from the axis in the author's program, which for each point of the point cloud calculates its distance from the axis and estimates the density of points per square meter of the grid. The results show that, despite more frequent measurements, the density of points dcreases with increasing speed at high speeds. The optimum driving speed (in relation to the number of points in the cloud) varies between 25 and 35 km/h. At this speed, it is possible to determine the pattern of horizontal road markings without error and to identify the technical means of traffic management on the road. The paper also presents performance measurements for these raw data sets, which will help researchers when analysing large volumes of material. The versatility of the proposed approach makes it independent of a specific type of scanner, which expands the possibilities of application. Using two types of data at once – point clouds and panoramic images – increases the accuracy of the information model created for the needs of road construction and operation. The point cloud provides spatial reference and assessment of the evenness of the surface, and panoramic images allow you to identify and model elements of the TSODD, such as signs, types of traffic lights and road markings.