Robot Apparatus 3D Position Estimation via 2D Contour Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing robot apparatuses face difficulties in accurately recognizing and holding minute objects, such as needles or pills, due to low accuracy in three-dimensional form measurement and complex operation instructions, which leads to ineffective holding operations.
Innovation Solution
A robot apparatus equipped with an output unit for displaying images, an input unit for user specifications of an approximate object range, an object extraction unit for extracting two-dimensional contours, and a position and attitude estimation unit for estimating three-dimensional positions, allowing for improved object recognition and holding operations through simpler user instructions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If three-dimensional form measurement is used to obtain object position data, then the robot apparatus can identify objects in the environment, but the measurement accuracy is low and insufficient for holding minute objects
Solution Approach 1:
The patent transitions from three-dimensional form measurement to two-dimensional image processing by capturing images from multiple angles and synthesizing them. This dimensional approach change enables higher precision positioning of minute objects by utilizing contour extraction and feature point detection in the two-dimensional image plane, thereby resolving the measurement accuracy insufficiency for holding operations.
Solution Approach 2:
The patent introduces an image processing system as an intermediary between the robot apparatus and the object. By capturing images with an image pickup device, extracting contours, and calculating positions based on image coordinates, the system achieves high-precision measurement without directly relying on low-accuracy three-dimensional form measurement, thus improving holding operation effectiveness.
2Adaptability or versatility
If a database of principal form models is used for object recognition, then the robot apparatus can identify common objects, but it cannot treat objects that do not match registered models
Solution Approach 1:
The patent enables the robot apparatus to automatically recognize objects without requiring pre-registered models or complex user operations. By capturing images from multiple angles, extracting contours automatically, and calculating positions based on image processing, the system achieves versatile object recognition for both common and novel objects, eliminating the need for database matching and simplifying the operation instruction system.
3Ease of operation
If the user specifies the object through the instruction terminal, then the robot apparatus can perform holding operations, but the operation process becomes complex and users may avoid using the system
Solution Approach 1:
The patent replaces the mechanical interaction of complex user operations with an automated image processing system. Instead of requiring users to manually specify objects through complex terminal operations, the system automatically captures images, extracts contours, identifies objects, and calculates positions, thereby simplifying user interaction while maintaining high measurement precision for holding operations.
Data Source
AI summary
A robot apparatus includes an output unit that displays an image including an object on a screen, an input unit that receives an operation performed by a user for specifying information relating to an approximate range including the object in the image, an object extraction unit that extracts information regarding a two-dimensional contour of the object on the basis of the specification received by the input unit, and a position and attitude estimation unit that estimates information regarding a three-dimensional position and attitude of the object on the basis of the information regarding the two-dimensional contour.