3D Scene Map Generation Using Point Cloud Data Registration and Model Replacement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for generating 3D static environments using point cloud data are costly and inefficient, requiring high manual effort or large amounts of data, limiting their application to other scenes due to high precision requirements and inconvenience in transporting laser scanners.
Innovation Solution
A method involving data registration of point cloud data from multiple frames, deletion of movable obstacle data, and replacement of regularly shaped object data with geometry model data to generate a 3D scene map, using a combination of global and local registration algorithms and 3D modeling techniques to create a realistic and efficient 3D background map.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If manual modeling method is used to create 3D grid model, then the model can be created with high precision, but the cost of manpower and material resources increases significantly
Solution Approach 1:
The patent uses copying by replacing manually created 3D grid models with automatically generated models from point cloud data. The system captures real-world scenes using laser scanners or cameras to obtain point cloud data, then automatically generates 3D grid models through processing algorithms, eliminating the need for manual modeling while preserving geometric accuracy.
Solution Approach 2:
The patent applies mechanics substitution by replacing the manual mechanical modeling process with an automated computational system. Instead of manual operations to create 3D models, the system uses point cloud processing algorithms including coordinate system transformations, grid generation, and automatic mesh creation to produce equivalent or superior results with minimal human intervention.
2Measurement precision
If high precision laser scanner is used to collect point cloud data, then the data quality is high, but the device complexity and transportation inconvenience increase
Solution Approach 1:
The patent applies segmentation by dividing the data collection process into multiple phases: rough scanning with simpler devices to capture overall scene geometry, followed by targeted high-precision scanning of specific regions of interest. This allows the system to achieve high overall data quality without requiring high-precision scanners to be deployed everywhere, reducing device complexity and transportation needs.
Solution Approach 2:
The patent implements universality by designing a processing system that can handle point cloud data from multiple device types and precision levels. The system processes data from both high-precision laser scanners and lower-precision cameras or scanners uniformly through the same coordinate transformation and grid generation algorithms, making the system adaptable to different device capabilities without requiring specialized equipment for each scenario.
3Reliability
If high density point cloud data is collected to ensure scene realism, then the simulation quality improves, but the quantity of data increases leading to higher processing costs
Solution Approach 1:
The patent applies local quality by varying the point cloud density according to the importance and complexity of different scene regions. The system identifies regions requiring high detail (such as obstacles, road features, or areas with complex geometry) and processes them with higher density sampling, while using lower density sampling in less critical areas. This maintains scene realism where needed while reducing overall data volume and processing requirements.
Solution Approach 2:
The patent implements partial action by collecting and processing only the necessary portion of point cloud data required to achieve the desired simulation quality. The system uses preliminary processing to identify essential scene elements and focuses high-density data collection on those specific elements rather than uniformly processing the entire scene, thereby reducing total data volume while maintaining realism in critical areas.
Data Source
Figure 1~2
Figure 3
Figure 4~5
AI summary
The present disclosure provides a method and a device for generating a 3D scene map, a related apparatus and a storage medium. The method includes the following. At least two frames of point cloud data collected by a collection device is obtained (101). Data registration is performed on the at least two frames of point cloud data. First type of point cloud data corresponding to a movable obstacle is deleted from each frame of point cloud data and each frame of point cloud data is merged to obtain an initial scene map (102). Second type of point cloud data corresponding to a regularly shaped object is replaced with model data of a geometry model matching with the regularly object for the initial scene map to obtain the 3D scene map (103).