Texture-Less Metal Part Pose Matching for Reliable Robotic Grasping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for grasping texture-less metal parts in computer vision and industrial automation face challenges due to low matching accuracy and sensitivity to external factors like contrast variations and pose changes, making it difficult to achieve reliable grasping.
Innovation Solution
A method using a novel line segment direction definition and improved distance function calculations within the Bundle Of Lines Descriptor (BOLD) framework, which involves extracting line segments, constructing descriptors based on angle relations, and using perspective n lines to match and recognize poses for accurate grasping of polyhedral metal parts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If BOLD descriptor is used for line segment matching, then matching accuracy is improved, but the method becomes sensitive to contrast changes causing mismatches
Solution Approach 1:
The patent changes the parameter representation from gradient magnitude-dependent BOLD descriptors to gradient direction-only descriptors. By removing the magnitude component that is sensitive to contrast changes and keeping only the directional information, the method maintains matching accuracy while becoming robust to contrast variations. This is achieved by redefining the line segment descriptor to use only angle relations between line segments and their nearest neighbors, eliminating contrast sensitivity.
2Ease of manufacture
If contour-based matching methods are used, then the approach is simple to implement, but matching accuracy is low and success rate is affected
Solution Approach 1:
The patent segments the contour into multiple line segments and further segments the description of each line segment into angular relationships with k nearest line segments. This hierarchical segmentation transforms the simple but inaccurate whole-contour comparison into a detailed line-by-line matching process, significantly improving matching accuracy while maintaining computational feasibility through localized descriptor construction.
3Productivity
If feature point extraction methods (SIFT or SURF) are used, then grasping efficiency is high and accuracy is good, but the method cannot be applied to texture-less metal parts
Solution Approach 1:
The patent extracts and utilizes the geometric structure information from line segments in the image, taking out the essential shape characteristics needed for matching. By focusing on the extracted line segment geometry and their angular relationships rather than texture features, the method achieves efficient and accurate matching specifically adapted for texture-less metal parts, replacing the need for texture-dependent feature points.
Data Source
AI summary
A method for grasping texture-less metal parts based on BOLD image matching comprises: obtaining a real image and CAD template images by photographing, extracting a foreground part of the input part image, calculating a covariance matrix of a foreground image, establishing the direction of a temporary coordinate system, and setting directions of line segments to point to a first or second quadrant of the temporary coordinate system; constructing a descriptor of each line segment according to an angle relation between the line segment and k nearest line segments, and matching the descriptors of different line segments in the real image and the CAD template images to obtain line segment pairs; and recognizing a processed pose through a PNL algorithm to obtain a pose of a real texture-less metal part, and then inputting the pose of the real texture-less metal part to a mechanical arm to grasp the part. The present invention can correctly match line segments, obtain an accurate pose of the part by calculation, successfully grasp the part, and satisfy actual application requirements.