Dynamic Point Cloud Filtering for 3D Road Surface Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for generating detailed maps for self-driving cars using mobile mapping systems face challenges in filtering out irrelevant point cloud data, such as buildings and vehicles, due to limitations in filtering technology, often requiring restrictive methods like nighttime data collection or fixed threshold values, which result in errors and do not consider environmental factors like view height and road gradient.
Innovation Solution
A method and apparatus that process point cloud data by determining the view height of a laser scanner and reference height based on GPS data, filtering point cloud data within specific height ranges using adjustable window sizes and threshold values, and generating three-dimensional road surfaces using convex hull algorithms and raster generation techniques.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If fixed threshold value filtering method is used, then filtering process is simple, but filtering accuracy is low and errors are present because environmental factors like view height and road gradient are not considered
Solution Approach 1:
The patent transforms the static fixed threshold filtering method into a dynamic adaptive filtering method. The threshold values for filtering point cloud data are no longer fixed but are dynamically adjusted based on environmental factors including view height of the laser scanner and road gradient. This allows the filtering system to adapt to different operating conditions, thereby improving filtering accuracy while maintaining process simplicity through automated parameter adjustment.
Solution Approach 2:
The patent changes the filtering parameters (threshold values) based on environmental conditions. Specifically, the view height of the laser scanner and road gradient are used to dynamically modify the threshold values applied during point cloud filtering. This parameter adaptation enables the system to accurately filter road surface points under varying environmental conditions, resolving the contradiction between simple filtering process and high filtering accuracy.
2Reliability
If restrictive methods like nighttime data collection are used, then filtering of irrelevant objects is easier, but productivity is reduced due to limited operational time
Solution Approach 1:
The patent replaces the mechanical restriction of nighttime-only data collection with an intelligent software-based filtering system. Instead of relying on environmental conditions (nighttime) to simplify filtering, the system uses advanced algorithms that automatically distinguish road surface points from irrelevant objects (buildings, vehicles, facilities) based on spatial relationships and geometric characteristics. This substitution enables high-quality data collection during both day and night, significantly improving productivity while maintaining filtering effectiveness.
Solution Approach 2:
The filtering system performs self-service by automatically identifying and separating road surface points from irrelevant objects without requiring manual intervention or restrictive operational conditions. The system uses the view height of the laser scanner and road gradient information to autonomously determine which points belong to the road surface, enabling continuous operation under various lighting conditions and improving overall data collection efficiency.
3Measurement precision
If manual extraction of road surface is used, then filtering accuracy is high, but device complexity and operation difficulty increase
Solution Approach 1:
The patent implements a self-service automated filtering system that performs road surface extraction without manual intervention. The system automatically uses the view height of the laser scanner and road gradient to identify and extract road surface points from the point cloud data. This automation maintains high extraction accuracy while eliminating the complexity and operational difficulty associated with manual methods, as the system performs all filtering operations autonomously based on pre-configured parameters.
Solution Approach 2:
The patent incorporates feedback mechanisms where the system continuously refines the road surface extraction process by using the determined view height and road gradient information to adjust filtering parameters in real-time. This feedback loop ensures high extraction accuracy while keeping the processing system relatively simple, as the automation handles the complexity of parameter adjustment and iterative refinement without requiring manual intervention.
Data Source
AI summary
Provided herein is a method for generating a road surface. The method for generating a road surface includes: obtaining a view height of a laser scanner used in an operation process through a mobile mapping system (MMS); determining a reference height on the basis of the obtained view height and a height measured by a global positioning system (GPS); extracting point cloud data positioned in a predetermined height range from the determined reference height among point cloud data obtained in the mobile mapping system; and generating the road surface on the basis of the extracted point cloud data.