The reconstruction of scanned objects from high-density point cloud data to extract the required data is an important part of data processing. According to the different ways of 3D model representation, model reconstruction of point cloud data can also be divided into two methods:

One is geometric model reconstruction, which is common in CAD models of contour models or section profiles. For geometric model reconstruction, mainly in the point cloud or in the data model, the object is fitted by geometry such as points, lines, faces, cylinders, tetrahedrons, and the like. One is model reconstruction of the surface of the 3D point cloud data, and the main structural grid (triangular patch) approaches the surface of the scanned object;

Commonly, such as geometric contouring of the outline of the building, the edge of the road, the layout of the wires, the pipe of the factory, etc., can output the three-dimensional graphic format or the AutoCAD format.

Accuracy analysis: Since the accuracy of the point position of 3D scanning depends on the error of distance measurement and angle measurement. According to statistics, the highest precision of the current non-reflective ranging is 3+2ppm, and the highest precision is 0.5 seconds. According to the error propagation law in the process of determining the position, the error of single point position determination in 3D scanning will be greater than 3mm, and as the distance from the scanner to the scanning target increases, the error determined by the point will be measured by distance and angle. The error increases and becomes larger. It can be seen that the current three-dimensional scanning technology can not achieve the mm level mapping accuracy.

When the scanning target phase is located 5 meters away from the scanner, the three-dimensional scanning result contains an error of about 3 to 5 mm (due to careful manual measurement, the error in the manual measurement result can be ignored). Although the relationship between the measurement error and the scanning distance is not obtained, according to the characteristics of the laser ranging and positioning, it can be concluded that as the distance from the scanning object to the scanner increases, the measurement error will increase greatly.

Since the point determination is related to the distance, the accuracy of the 3D laser scanning is not a fixed value. The accuracy of ranging is affected by the length of measurement and the number of measurements. The farther the scanner is from the target, the greater the error in distance measurement and the lower the accuracy of the point. The greater the number of ranging, the higher the accuracy of ranging. The purpose of 3D laser scanning is to determine the position of the point. Therefore, like other measuring devices, the index for measuring the accuracy of 3D laser scanners is single point positioning accuracy. According to the principle of determining the position in the 3D scanner, the accuracy of the point measurement depends on the accuracy of the instrument and the accuracy of the angle measurement.

There are many factors affecting the accuracy of laser scanning measurement. In general, there are two aspects of hardware and software. The hardware mainly includes mechanical motion platform, CCD camera, laser, etc. The accuracy is improved from the hardware aspect, which generally increases the cost. The software mainly includes the object image correspondence calibration, laser scanning line center extraction, measured object surface features, optical imaging parameters, light plane position and other factors.

to sum up:

3D laser scanning technology is an emerging technology that has been developed in recent years and has a wide range of applications. The emergence and continuous development and promotion of 3D laser scanning technology has brought new working methods to the acquisition of geoscience information and is a useful supplement to traditional geoscience survey methods. The 3D laser scanning technology has absolute advanced technology advantages, and its application to engineering measurement and deformation monitoring in geotechnical and geological engineering fields has great application potential.

Through the introduction and analysis of this paper, the 3D laser scanning technology has the advantages of long measuring distance, fast sampling point, high point positioning accuracy, no contact measurement (no need for reflective prism), scanning target without surface processing, and directly acquiring 3D point cloud data, and simple data processing. Fast, software powerful and comprehensive, can basically meet the needs of engineering measurement.

In recent years, with the deepening of the research and application of laser scanning technology, the research on 3D reconstruction technology based on laser scanning data has received extensive attention, and some research results have been obtained at home and abroad. The goal of building 3D model reconstruction based on laser scanning data is to obtain a 3D model of the scene with a more accurate 3D geometry. This paper discusses data acquisition, point cloud processing and measurement accuracy analysis of building 3D models based on laser scanning data. Through the analysis and comparison of data processing and results, the conclusions on the accuracy of 3D laser scanning are obtained.