3D Scene Capture With Pose Uncertainty Feedback for SLAM Drift
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing SLAM systems for scanning structural objects suffer from increasing spatial error due to drift, leading to inaccuracies in composite point clouds, especially when using mobile scanners, and there is no effective way to correct this in a user-aware manner.
Innovation Solution
A method that propagates uncertainty in the pose graph of a mobile scanning system, allowing users to visually identify regions of high spatial error and take corrective actions in real-time by revisiting areas with high uncertainty, using a mobile scanning device with a Lidar sensor and IMU, and displaying point clouds with color-coded uncertainty levels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If mobile scanning with SLAM is used to enable real-time scanning and mapping, then scanning speed and productivity are improved, but spatial error increases due to drift
Solution Approach 1:
The system calculates uncertainty values for each location in the pose graph and provides visual feedback through color-coded indicators on the display. This allows users to see which areas have high spatial error and need revisiting, enabling real-time correction of drift accumulation without stopping the scanning process.
Solution Approach 2:
The system proactively identifies regions with high uncertainty before the scanning is complete, allowing users to plan corrective revisits in advance. The uncertainty propagation calculates potential error accumulation ahead of time, enabling preventive action rather than post-processing correction.
2Measurement precision
If uncertainty calculation and visualization are added to help users identify high error regions, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The system introduces an uncertainty propagation module as an intermediary between the SLAM system and the user interface. This module calculates uncertainty values based on existing pose graph data and translates complex error accumulation into simple visual indicators, adding precision without significantly complicating the core scanning system.
Solution Approach 2:
The system uses color-coded indicators to represent different levels of uncertainty in the pose graph. This visual encoding transforms complex spatial error data into intuitive color signals that users can quickly interpret, improving measurement precision identification without adding complex interface elements.
3Measurement precision
If users revisit areas with high uncertainty to reduce drift, then spatial error is reduced, but loss of time increases due to additional scanning required
Solution Approach 1:
The system calculates and displays uncertainty values throughout the scanning process, allowing users to identify high-error regions and plan efficient revisit routes before completing the scan. This preliminary identification enables users to optimize their revisiting strategy and minimize redundant scanning time.
Solution Approach 2:
The system automatically identifies which areas need revisiting and provides guidance to users, enabling them to self-correct drift accumulation without external assistance. The uncertainty propagation and visual indication system serves itself by highlighting problem areas, reducing the need for time-consuming manual quality checks.
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
Enables users to reduce spatial error in point clouds by identifying and addressing areas of high uncertainty during scanning, improving the accuracy of composite point clouds without requiring extensive computational effort, thus enhancing the quality of scanning results.
Implementation Method 1
The system comprises a mobile scanning device configured to scan the structural object from a plurality of successive locations to generate for each location a respective point cloud
Implementation Method 2
Often Lidar scanners also include IMUs (inertial measurement units) and/or cameras
Data Source
AI summary
A system is disclosed for scanning a structural object, comprising a mobile scanning device configured to scan the structural object from a plurality of successive locations to generate for each location a respective point cloud representing a respective portion of the structural object; a calculating unit configured to: (i) determine from the point clouds a pose graph comprising estimates of positions of the successive locations in relation to the structural object; and (ii) calculate for each location a respective uncertainty value in the pose graph; and a display configured to display the point clouds and to provide a visual indication for each point cloud of the calculated uncertainty value of the corresponding location in the pose graph.