Semantic Landmark Scanning for Accurate Environment Registration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing environment scanning systems face challenges in automating the registration of multiple scans, leading to inefficiencies and inaccuracies, particularly in large-scale scanning applications, due to the reliance on manual processes and the inability to accurately match and align scans without sufficient overlap or proper registration of flat surfaces.
Innovation Solution
A system that employs a mobile scanning platform with a scanner device capable of simultaneous localization and mapping (SLAM) using semantic features, such as corners and measurable walls, to automatically detect and utilize landmarks for accurate registration and alignment of scans, reducing the need for manual intervention and improving scan accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual registration processes are used to align multiple scans, then scan accuracy can be maintained through human judgment, but productivity is reduced due to time-consuming manual operations
Solution Approach 1:
The system performs automatic scan registration using semantic feature extraction and matching algorithms that enable the scanner to self-align multiple scans without manual intervention. The processor automatically identifies semantic features in overlapping scan areas, computes transformation parameters, and aligns scans based on these computed features, making the registration process autonomous and significantly faster while maintaining accuracy.
Solution Approach 2:
The patent replaces manual mechanical registration operations with automated computational processes. Instead of human operators manually aligning scans, the system uses processors to execute algorithms that extract semantic features, match features between scans, and compute alignment transformations automatically, substituting human judgment and manual manipulation with automated image processing and computational geometry.
2Productivity
If automated registration is implemented without semantic features, then productivity increases through automation, but measurement precision deteriorates due to inability to accurately match flat surfaces
Solution Approach 1:
The system applies different feature extraction strategies to different regions of the scan data. In areas with rich geometric details, traditional feature matching is used, while in flat surface regions, semantic features such as detected corners, lines, and structural elements are specifically targeted. This localized approach ensures accurate matching in challenging flat areas while maintaining overall automation and productivity.
Solution Approach 2:
The patent transforms scan data from raw coordinate information into semantic feature representations by detecting corners, lines, and structural elements. This parameter transformation converts ordinary point cloud data into meaningful geometric primitives that provide distinctive markers for accurate feature matching, enabling precise alignment even on previously problematic flat surfaces while maintaining automated operation.
3Measurement precision
If multiple scans are performed to cover large areas, then measurement precision improves through comprehensive coverage, but loss of time increases due to the number of scans required
Solution Approach 1:
The system performs preliminary processing of each scan immediately upon acquisition, extracting and storing semantic features as the scan is being captured. This preliminary feature extraction prepares the data for rapid matching and alignment, reducing the computational burden during the registration phase and enabling faster processing of multiple scans without compromising the comprehensive coverage needed for complete environmental mapping.
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 efficient, accurate, and autonomous scanning and mapping of environments by automatically identifying and utilizing semantic features as landmarks, thereby reducing errors and increasing the speed and precision of scan registration, even in complex or time-sensitive scenarios.
Implementation Method 1
A TOF laser scanner is a scanner in which the distance to a target point is determined based on the speed of light in air between the scanner and a target point
Implementation Method 2
The beam steering mechanism includes a first motor that steers the beam of light about a first axis by a first angle that is measured by a first angular encoder
Data Source
AI summary
A method is provided that includes recording a landmark at a first scan position of a scanner, the landmark based at least in part on a semantic feature of scan data captured by the scanner. The semantic feature is identified using line-segments of the scan data. The method further includes capturing, by the scanner while moving through the environment, additional scan data at a second scan position. The method further includes, responsive to the scanner returning to the first scan position associated with the landmark, computing a measurement error. The method further includes correcting, using the measurement error, at least a portion of the scan data or the additional scan data.