3D Point Cloud Position Matching Using Distance Histograms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High accuracy estimation of a mobile body's position is challenging when using three-dimensional point cloud data with different densities, as low-density data from a mobile body sensor is difficult to associate with high-density map data.
Innovation Solution
A database is generated that registers combinations of reference points and feature amounts in a mobile space, allowing for the extraction of reference points based on distance calculations from three-dimensional point data, enabling association of point clouds with different densities by calculating histograms for each reference point and using these to specify corresponding positions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If three-dimensional point cloud data with different densities is used for position estimation, then the mobile body can operate with limited sensor data, but the alignment accuracy between map data and sensor data degrades
Solution Approach 1:
The patent transforms the point cloud data from its original three-dimensional coordinate representation into a two-dimensional histogram representation based on distance frequencies. This parameter transformation allows low-density sensor data to be effectively compared with high-density map data by changing the representation space, thereby resolving the alignment accuracy degradation issue while maintaining operability with limited sensor data
Solution Approach 2:
The patent introduces histograms as an intermediary representation between the raw point cloud data and the position estimation process. By converting both map data and sensor data into histogram form based on distance frequencies from reference points, the system creates a common intermediate representation that enables accurate alignment despite density differences, thus resolving the contradiction between adaptability and measurement precision
2Measurement precision
If high-density map data is used for accurate position estimation, then position accuracy improves, but the system becomes dependent on pre-existing high-density data
Solution Approach 1:
The patent performs preliminary conversion of the point cloud data into histogram representations before the actual position estimation process. By pre-calculating the distance frequency distributions from reference points and storing them as histograms, the system prepares the data in a form that can be directly compared and matched, enabling accurate position estimation without requiring the data to remain in its original high-density format
Solution Approach 2:
The patent creates a simplified copy of the point cloud data in the form of histograms that capture the essential spatial distribution characteristics without preserving the full high-density detail. This histogram copy retains the necessary information for accurate position estimation while being independent of the original high-density data structure, thus reducing dependence on pre-existing high-density data while maintaining position accuracy
Data Source
AI summary
A recording medium records a program that causes a computer to execute processing of: obtaining pieces of three-dimensional point data observed a mobile body; calculating a feature amount according to a distance between a position of the mobile body and each of three-dimensional points indicated by the pieces of three-dimensional point data; and specifying a reference point that corresponds to the position of the mobile body from among reference points in a mobile space of the mobile body which are registered in a database that registers a combinations each of the respective reference points and the feature amounts according to the distance between the position of the respective reference points and each of three-dimensional points indicated by a pieces of three-dimensional point data observed from the position of the respective reference points, based on the calculated feature amount and the database.