Robot Gripping Position Detection Using Repeating Image Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining the gripping position of a target object by a robot are either time-consuming due to the need for teaching models or lack accuracy when not using them, especially when the target objects are randomly piled up or changed.
Innovation Solution
An information processing apparatus that acquires images of the target object, detects image features, clusters them into similar groups, and calculates the gripping position based on the gravity center of these groups, allowing for accurate and rapid determination without relying on teaching models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If teaching model data is used to estimate gripping position, then estimation accuracy is improved, but preparation time increases
Solution Approach 1:
The patent uses image information as a copy or representation of the target object's physical properties. By extracting features from images rather than using physical teaching models, the system achieves accurate gripping position estimation without the time-consuming process of creating and storing teaching model data for each object type.
Solution Approach 2:
The patent replaces the mechanical approach of physical teaching models with an information processing approach using image features and algorithms. The gripping position is determined through computational analysis of image data rather than through physical measurement and model creation, significantly reducing preparation time while maintaining accuracy.
2Measurement precision
If teaching model is generated for each target object change, then gripping accuracy is maintained, but work efficiency decreases
Solution Approach 1:
The patent creates a universal system that can handle different target objects using the same image processing pipeline and feature extraction algorithms. Instead of requiring separate teaching models for each object type, the system universally applies image-based feature detection and gripping position calculation, enabling rapid adaptation to different objects without retraining or model regeneration.
Solution Approach 2:
The patent performs preliminary feature extraction and analysis on image data to identify gripping positions before actual gripping operations begin. By pre-processing image information and extracting relevant features in advance, the system prepares gripping position data rapidly, improving work efficiency while maintaining accuracy across different object types.
3Loss of time
If no teaching model is used, then preparation time is reduced, but estimation accuracy decreases
Solution Approach 1:
The patent enables the system to automatically extract features and determine gripping positions directly from image information without requiring external teaching model data. The image processing system serves itself by using the target object's own visual features to determine gripping positions, eliminating the need for separate model creation while maintaining high accuracy through sophisticated feature extraction algorithms.
Data Source
AI summary
An information processing apparatus that obtains, with high accuracy, the gripping position for gripping the target object by a gripping apparatus is provided. The information processing apparatus includes an image acquisition unit that acquires a captured image obtained by capturing an image of the target object, an image feature detection unit that detects a plurality of image features in the captured image, and a unit detection unit that detects a unit of a pattern in which a plurality of image features repeatedly appear. In addition, the information processing apparatus further includes a snipping position determining unit that determines a gripping position using the unit detected by the unit detection unit.