Point Cloud Clustering for Clear Crane Guide Frames
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for displaying guide information on a monitor for measurement target objects, such as buildings, can become overly complex and reduce visibility when dealing with objects of complex shapes or multiple objects in close proximity, leading to excessive information overload.
Innovation Solution
A clustering method for point cloud data that uses a laser scanner to acquire data from measurement target objects, processing it to identify and group planar clusters based on elevation differences and overlap detection, creating a guide frame that collectively encloses clusters to enhance visibility on a monitor.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If conventional three-dimensional point cloud analysis methods are used to display guide information for each measurement target object, then the outer shape and height information of buildings can be displayed, but the information becomes excessive and visibility is reduced when dealing with complex shapes or multiple objects in close proximity
Solution Approach 1:
The patent segments the point cloud data by extracting top surface points and forming planar clusters, then groups these clusters into super-clusters representing individual measurement target objects. This segmentation approach separates the top surface information from other point cloud data, reducing information overload while maintaining visibility of guide information for each object.
2Measurement precision
If detailed point cloud data of multiple objects is displayed individually, then comprehensive shape information is provided, but the complexity of the display increases and reduces clarity
Solution Approach 1:
The patent extracts only the top surface points from the complete point cloud data by identifying points with upward normal vectors. This extraction provides sufficient shape and height information for guide display while significantly reducing display complexity, as only essential top surface characteristics are shown rather than complete three-dimensional point cloud details.
3Reliability
If complete point cloud data is processed to identify all objects, then accurate object detection is achieved, but the processing time and computational load increase
Solution Approach 1:
The patent performs preliminary action by first extracting top surface points and forming planar clusters before proceeding to super-cluster formation and object identification. This preliminary extraction of essential top surface information reduces the data volume for subsequent processing steps, maintaining object detection accuracy while reducing overall processing time.
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
The method effectively clusters point cloud data to reduce information overload, improving visibility and clarity of guide information displayed on a monitor for users, particularly when dealing with complex shapes or multiple objects in the same region.
Implementation Method 1
point cloud data of a measurement target object acquired by a laser scanner
Implementation Method 2
acquiring a three-dimensional shape of a measurement target object on the basis of point cloud data of the measurement target object acquired by a laser scanner
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Provided is a method for clustering the data point groups of one or more measurement targets located in the same region from among the acquired data point groups. This method is provided with: a data point group acquisition step for acquiring a data point group in a region that contains a measurement target from above the measurement target by using a laser scanner; a clustering step for clustering the data point groups that correspond to the top surface of the measurement target as a planar cluster by using a data processing unit; a step for extracting a reference planar cluster which is a reference for making a same-region determination; a step for calculating the difference in height between the reference planar cluster and other planar clusters, and searching for planar clusters exhibiting a height difference within a prescribed threshold; a step for selecting one planar cluster exhibiting a height difference within the prescribed threshold; a step for detecting whether there is overlap between the reference planar cluster and the one planar cluster; and a step for clustering the planar clusters as clusters in the same region when overlap is detected.