3D Cable Model Generation from Point Cloud Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques face difficulties in creating three-dimensional models of cables using point cloud data from three-dimensional laser measuring equipment, as cables have shapes different from utility poles and require efficient two-dimensional diagnosis.
Innovation Solution
A detection apparatus and method that includes a point cloud data input unit, a rule-based three-dimensional model generation unit, a machine learning-based three-dimensional model generation unit, and a three-dimensional model merging unit to combine and generate three-dimensional models of line-like structures from point clouds, allowing for the creation of accurate cable models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If three-dimensional laser measuring equipment is used to create models of utility poles, then columnar structures can be accurately modeled, but line-like structures such as cables cannot be effectively modeled
Solution Approach 1:
The patent segments the modeling process into two distinct pathways: rule-based modeling for line-like structures and machine learning-based modeling for columnar structures. This segmentation allows each method to be optimized for its specific structure type, resolving the contradiction between modeling accuracy and adaptability to different structures.
Solution Approach 2:
The patent introduces a dual-path processing architecture that adds a methodological dimension to the modeling system. By creating separate processing streams for different structure types (line-like vs. columnar), the system achieves both high accuracy for each specific type and broad adaptability across multiple structure categories.
2Device complexity
If a single modeling method is used for both cables and utility poles, then the system is simple, but the modeling accuracy for cables deteriorates
Solution Approach 1:
The patent implements a dynamic modeling system that automatically selects the appropriate modeling method based on the input structure type. The system dynamically adjusts its processing approach - using rule-based methods for line-like structures and machine learning for columnar structures - thereby maintaining high accuracy for cables while managing system complexity through automated decision-making.
Solution Approach 2:
The patent changes the methodological parameters of the modeling system based on the target structure. By detecting whether the target is line-like or columnar and switching between different modeling algorithms, the system achieves high cable modeling accuracy without requiring manual intervention, thus balancing complexity and precision.
3Reliability
If visual inspection by maintenance workers is used, then facility states can be assessed, but the process is time-consuming and inefficient
Solution Approach 1:
The patent replaces the manual visual inspection mechanism with an automated three-dimensional modeling system. By substituting human workers with laser measuring equipment and automated processing algorithms, the system maintains reliable diagnosis capability while dramatically improving inspection efficiency and productivity.
Solution Approach 2:
The patent creates three-dimensional digital copies of utility poles and cables from point cloud data. These digital models serve as accurate representations that can be analyzed automatically, replacing the need for physical visual inspection while preserving diagnostic reliability and enabling efficient automated analysis.
Data Source
AI summary
An object of the present disclosure is to provide a technique for creating a three-dimensional model of a line-like structure from a point cloud obtained using three-dimensional laser measuring equipment and detecting a three-dimensional model of a cable. A detection apparatus according to the disclosure includes a point cloud data input unit 12 that reads point cloud data where a structure that is present in a three-dimensional space is represented by a point cloud that is present in the three-dimensional space, a rule-based three-dimensional model generation unit 15 that combines linearly disposed point clouds into a group and generates a three-dimensional model of a line-like structure using a direction vector configured with point clouds included in the group, a machine learning-based three-dimensional model generation unit 14 that generates a three-dimensional model of a line-like structure based on a database that links point clouds and line-like structures, and a three-dimensional model merging unit that selects one of a plurality of three-dimensional models of line-like structures generated at an identical position in the three-dimensional space as a three-dimensional model of a line-like structure that is present in the three-dimensional space and merges three-dimensional models of the line-like structures that are present in the three-dimensional space.