Ground Point Cloud Extraction Method Combining Multi-Level Progressive Strategies and Unsupervised Learning

Abstract

The invention discloses a ground point cloud extraction method that combines multi-level progressive strategies and unsupervised learning, comprising the following steps: S1, acquiring point cloud data and performing noise removal; S2, adopting a Gaussian mixture model to realize the point cloud automatic clustering method ; S3, automatically acquire and mark the training samples of ground points and object points; S4, calculate the point cloud feature vector and use the SVM method to establish the training model, and use the trained SVM model to classify the LiDAR point cloud; S5, use the confidence interval The estimation theory automatically identifies and eliminates non-ground points; S6, repeating steps S2 to S5 until there is no ground point in the point cloud. The invention adopts a multi-level progressive strategy to realize the change of ground point cloud extraction results from coarse to fine, enhance the robustness of the method, realize automatic ground point cloud extraction, and avoid complicated parameter adjustment.

Description

technical field [0001] The invention relates to the technical field of geospatial information systems, in particular to a ground point cloud extraction method that integrates multi-level progressive strategies and unsupervised learning. Background technique [0002] With the rapid development of smart cities, there is an urgent need for us to obtain the surrounding terrain and feature information in a more timely, accurate and efficient manner. In recent years, airborne LiDAR (Light Detection And Ranging) technology has provided us with a new observation method to obtain high-temporal-spatial resolution earth spatial information. Airborne LiDAR technology adopts the method of active measurement, which can quickly and accurately obtain the three-dimensional coordinate information of objects, and has the characteristics of fast speed and high precision. Moreover, the technology is not affected by the external environment such as light and light and shade changes, and can coll...

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to solve the problems of low precision, poor adaptability to complex terrain environments and the need for complex parameter adjustments in the process of ground point cloud extraction, and propose a fusion of multi-level progressive strategies and unsupervised learning methods for automation Ground Point Cloud Extraction Method

Images