3D Point Cloud Object Separation Using Outlier and Plane Removal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for recognizing objects in three-dimensional point clouds face challenges in accurately separating individual objects from the cloud data, particularly due to the presence of outliers and plane surface areas, which affects the speed and accuracy of object recognition in robotic systems.
Innovation Solution
A method that involves acquiring a three-dimensional point cloud image, eliminating outliers by downsampling and determining distance-based criteria, removing plane surface areas by identifying optimal planes, and clustering points by assigning labels to improve object separation and recognition precision.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If clustering is performed on the entire three dimensional point cloud image, then complete object information is captured, but processing time increases significantly and accuracy decreases due to outliers and plane surfaces
Solution Approach 1:
The patent segments the point cloud data into three distinct categories: outlier points, plane surface points, and object points. This segmentation is achieved through statistical analysis of point distribution characteristics, allowing the system to process only relevant object points for clustering, thereby reducing processing time while maintaining accuracy.
Solution Approach 2:
The patent performs preliminary elimination of outliers and plane surface points before the main clustering operation. By pre-processing the data to remove irrelevant points, the subsequent clustering operation works on a reduced dataset, significantly decreasing processing time without compromising object recognition accuracy.
2Productivity
If outliers are eliminated by downsampling, then processing speed improves, but object recognition accuracy may deteriorate due to loss of detail information
Solution Approach 1:
The patent applies different processing strategies to different regions of the point cloud based on their local characteristics. Outlier points are identified and eliminated using statistical methods that preserve genuine object features while removing spurious data. This localized quality-based filtering maintains processing speed while preserving recognition accuracy.
3Measurement precision
If plane surface areas are eliminated, then object separation from background improves, but complexity of the elimination process increases
Solution Approach 1:
The patent transforms the plane surface elimination problem into a parameter-based filtering task. By analyzing the geometric parameters of point distributions (such as coplanarity and spatial arrangement), the system automatically identifies and eliminates plane surface areas. This parameter-driven approach improves object separation precision while managing process complexity through mathematical formulation.
Data Source
Figure 1
Figure 2a
Figure 2b
AI summary
A method of separating an object in a three dimensional point cloud including acquiring a three dimensional point cloud image of an object using an image acquirer, eliminating an outlier from the three dimensional point cloud image using a controller, eliminating a plane surface area from the three dimensional point cloud image, of which the outlier has been eliminated using the controller, and clustering points of an individual object from the three dimensional point cloud image, of which the plane surface area has been eliminated using the controller.