Spaceborne lidar error prediction method based on geographic probes and machine learning

By constructing a nonlinear elevation error prediction model based on geographic detectors and machine learning, the accuracy problem of ICESat-2ATL08 data under complex terrain conditions was solved, enabling accurate prediction of elevation errors and automated selection of ground control points, thus improving data quality.

CN122131279APending Publication Date: 2026-06-02青海省基础测绘院
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
青海省基础测绘院
Filing Date
2026-02-11
Publication Date
2026-06-02

AI Technical Summary

Technical Problem

The accuracy of ICESat-2ATL08 elevation data decreases significantly under complex terrain conditions. Traditional error assessment methods lack analysis of the comprehensive influence of multiple factors, making it difficult to accurately identify and quantify the key factors affecting elevation error and their interactions. Furthermore, there is a lack of reliable methods to select suitable ground control points.

Method used

A method based on geographic detectors and machine learning is adopted. Through data analysis and feature extraction, combined with high-resolution optical satellite imagery and reference digital elevation models, the geographic detector model is used to conduct factor detection and interactive detection, construct a nonlinear elevation error prediction model, screen significant influencing factors, and use machine learning algorithms to train the model.

Benefits of technology

It enables accurate prediction of elevation errors of spaceborne lidar, provides a high-quality basis for data quality classification, supports automated screening of high-precision ground control points, and improves data quality in complex terrain areas.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122131279A_ABST
    Figure CN122131279A_ABST
Patent Text Reader

Abstract

This invention provides a method for predicting errors in spaceborne lidar based on a geographic detector and machine learning, belonging to the fields of remote sensing data processing and geographic information technology. The method involves: obtaining multi-dimensional feature factors including terrain parameters and quality indicators based on spaceborne lidar data; obtaining a reference digital elevation model based on high-resolution optical satellite stereo imagery; obtaining elevation error data using a difference calculation method based on the terrain elevation values ​​in the spaceborne lidar data; analyzing the multi-dimensional feature factors using a geographic detector model to obtain significant influencing factors; constructing and training a model based on the elevation error data to obtain an elevation error prediction model; and calculating the prediction error value for each footpoint based on the spaceborne lidar footpoint data. This invention solves the problems of unstable accuracy of existing elevation data under complex terrain conditions, incomplete analysis of error influencing factors, and insufficient error prediction capabilities.
Need to check novelty before this filing date? Find Prior Art