Laser point cloud data-based road single component automatic modeling method, system and device, and storage medium

Abstract

The invention discloses a laser point cloud data-based road single component automatic modeling method, system and device and a storage medium, and the method comprises the following steps: obtainingroad data which comprises point cloud data and image data; carrying out semantic segmentation on the road data to obtain vector data corresponding to the road component; and automatically generating acomponent model according to the vector data. The method has the beneficial effects that the vector of the component is automatically extracted, the extraction efficiency of the vector is high, the production period is short, the labor cost is low, automatic modeling is achieved, and the modeling efficiency is high.

Description

technical field [0001] The invention relates to modeling technology in the construction industry, in particular to an automatic modeling method, system, device and storage medium for road single components based on laser point cloud data. Background technique [0002] At present, the single data modeling of road facilities mainly adopts the methods of artificial blueprint flipping and oblique photographic modeling, etc., which has problems such as poor accuracy, low efficiency, and high cost, and cannot fully utilize the true effectiveness of new technologies. A large number of facilities The demand for intelligent maintenance is suppressed, so it is urgent to innovate technical means to promote industrial upgrading. [0003] Component modeling based on point cloud laser technology needs to extract the vector of the object to use for the corresponding 3D model of the component. At present, vector extraction is mostly manual selection, which requires a lot of labor costs. At ...

AI Technical Summary

Problems solved by technology

[0002] At present, the single data modeling of road facilities mainly adopts the methods of artificial blueprint flipping and oblique photographic modeling, etc., which has problems such as poor accuracy, low efficiency, and high cost, and cannot fully utilize the true effectiveness of new technologies. A large number of facilities The demand for intelligent maintenance is suppressed, so it is urgent to innovate technical means to promote industrial upgrading

[0003] Component modeling based on point cloud laser technology needs to extract the vector of the object to use for the corresponding 3D model of the component. At present, vector extraction is mostly manual selection, which requires a lot of labor costs. At the same time, the production cycle is long and the production efficiency is low.

Some studies have proposed automatic vector extraction methods for this process, but most of the methods are based on the extraction of certain types or types of features of components, which require certain prior knowledge, poor recognition of different intensities, and most of the techniques Although for a single facility component, there is no automatic vector extraction method. At the same time, the success rate of automatic vector extraction at the current stage is not high. In engineering practice, manual verification is required for the wrong part, so it is not widely used. Reduce costs and vector production cycle time

As for the modeling part, the current method cannot make full use of the collected point cloud data and image data, and the unstructured whole model built can hardly meet the later business needs, and lacks structured 3D road information

Images