Bare Earth Feature Extraction for 3D LiDAR Point Clouds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Analysis of 3D LiDAR point clouds in heavily forested areas is time-intensive and prone to missing features like trails or buildings due to manual foliage identification and removal.
Innovation Solution
A bare earth and feature extraction method that includes converting 3D point clouds into 2D images, identifying ground points through modified smoothing and histogramming, generating an inverse Above Ground Layer (AGL) to highlight ground features, and removing buildings using an 8-pass algorithm, thereby improving detection accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual foliage identification and removal is used to analyze 3D LiDAR point clouds, then the analyst can search for features below the foliage canopy, but the analysis process becomes time-intensive and features may be missed
Solution Approach 1:
The system performs automatic ground point classification and foliage removal without requiring manual analyst intervention. The algorithm independently processes the 3D LiDAR point cloud data, identifies ground points, and generates the inverse AGL product, making the system self-sufficient and eliminating time-consuming manual operations while maintaining consistent processing quality
Solution Approach 2:
The patent replaces the manual mechanical process of foliage identification and removal with an automated computational algorithm. The system uses point cloud processing techniques, including localized histogramming and spatial analysis, to automatically distinguish ground points from foliage returns, substituting human labor with machine-based processing that is both faster and more consistent
2Reliability
If manual foliage removal is performed to identify ground features, then features below canopy may be found, but the process is prone to human error and missed detections
Solution Approach 1:
The patent segments the LiDAR point cloud data into distinct categories: ground points, foliage returns, and above-ground features. By dividing the complex point cloud into manageable segments and applying specific processing rules to each, the system reliably identifies ground features while maintaining organized, systematic processing that reduces errors
Solution Approach 2:
The system introduces an intermediate processing product called the inverse Above Ground Level (AGL) product. This intermediary representation transforms the raw point cloud data into a format where ground features are emphasized and foliage is suppressed, serving as a bridge between raw data and final feature identification, thereby improving detection reliability
3Ease of operation
If 3D point clouds are converted to 2D images for analysis, then ground features become visually representable, but information dimensionality is reduced
Solution Approach 1:
The patent strategically transforms selected ground point data from 3D space to a 2D inverse AGL representation, creating a new dimensional perspective that emphasizes vertical variations in ground surface. This dimensionality change makes subtle ground features more visually apparent while the underlying system preserves access to the full 3D point cloud data for comprehensive spatial analysis when needed
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This method enhances the detection of human activity and hidden features by visually representing ground and above-ground elements, allowing for more accurate identification of paths, roads, and structures within dense foliage.
Implementation Method 1
Multiple 3D points may be collected using a LiDAR ranging system
Data Source
AI summary
Generally discussed herein are systems and apparatuses that relate to systems and methods for analysis of 3D light radar (LiDAR) point clouds. The disclosure also includes techniques of bare earth and feature extraction. According to an example, bare earth and feature extraction can include estimation of a ground layer, generation of an inverse Above Ground Layer (AGL), identification of paths, roads, and cleared areas, identification and removal of buildings, and identification of a Human Activity Layer (HAL). This combination of features may provide for a much higher probability of detection of human activity.